# Edge Impulse Documentation > Where you can find all the Edge Impulse documentation, and more! Edge Impulse is a platform for building, training, optimizing, and deploying machine learning models that run on edge devices - from microcontrollers to CPUs, GPUs, and NPUs. The documentation covers the full workflow: data collection, DSP and feature processing, model training and evaluation, hardware-aware optimization, on-device deployment, and lifecycle management, plus the Studio UI, SDKs, CLIs, and APIs used to automate it. This file is an llms.txt index of the Edge Impulse documentation, following the specification at https://llmstxt.org/. This is the curated `llms-full.txt` companion. It includes the complete authored-page inventory while intentionally collapsing generated Python model stubs and omitting generated OpenAPI endpoint pages. Use `llms.txt` to route through smaller section indexes. Links point to the markdown-rendered version of each page (`.md`), so agents can retrieve clean content without navigation, ads, or scripts. Every link includes a short note describing the page. Use the section headings to scope retrieval before expanding into individual pages. ## Intents Intent-bearing retrieval prompts. Match a user request to the closest intent below, then retrieve the linked pages before answering. ### Get started with Edge Impulse from scratch > I am new to Edge Impulse and want to understand the end-to-end workflow - collect data, design an impulse, train a model, and deploy it - so I can build my first edge AI project. - [Build a first project with guided, beginner-friendly explanations.](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-beginners.md) - [Configure the input, processing, and learning blocks that make up an impulse.](https://docs.edgeimpulse.com/studio/projects/impulse-design.md) - [Pick a complete worked tutorial that matches the project you want to build.](https://docs.edgeimpulse.com/tutorials/end-to-end/overview.md) - [Learn the core edge-AI concepts behind the workflow before building.](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals.md) ### Get started as an embedded engineer > I come from firmware and hardware and want to start from devices, resource constraints, firmware, and on-device deployment rather than from data science. - [Start from devices, firmware, resource constraints, and deployment.](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-embedded-engineers.md) - [Choose a supported board, device, or accelerator to target.](https://docs.edgeimpulse.com/hardware.md) - [Build a C++ library, firmware, package, or container for a specific target.](https://docs.edgeimpulse.com/studio/projects/deployment.md) ### Get started as an ML practitioner > I come from machine learning and want to start from datasets, model workflows, experiments, SDKs, and evaluation rather than from hardware. - [Start from datasets, model workflows, experiments, SDKs, and evaluation.](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-ml-practitioners.md) - [Compare alternative impulse and model configurations side by side.](https://docs.edgeimpulse.com/studio/projects/experiments.md) - [Evaluate trained models on held-out data before deploying.](https://docs.edgeimpulse.com/studio/projects/model-testing.md) - [Drive Studio workflows programmatically from the Python SDK.](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python.md) ### Train an image classification model and deploy it to a device > I want to collect labeled images, train a transfer-learning image classifier in Edge Impulse Studio, and deploy the model to a microcontroller or single-board computer. - [Follow an end-to-end image classification project from data to deployment.](https://docs.edgeimpulse.com/tutorials/end-to-end/image-classification.md) - [Collect and label image data directly in Studio.](https://docs.edgeimpulse.com/tutorials/topics/data/collect-image-data-studio.md) - [Train an image classifier with transfer learning on limited data.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/transfer-learning-images.md) - [Export the trained impulse to run on your target device.](https://docs.edgeimpulse.com/studio/projects/deployment.md) ### Detect, locate, and count objects in a camera feed > I want to build an object detection model that finds and counts multiple objects in real time on an embedded device, using FOMO or bounding boxes. - [Follow an end-to-end object detection and counting project.](https://docs.edgeimpulse.com/tutorials/end-to-end/object-detection-centroids.md) - [Use FOMO for fast, low-memory object detection on constrained devices.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection/fomo.md) - [Compare object detection block options and when to use each.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection.md) - [Count and track detected objects with post-processing.](https://docs.edgeimpulse.com/tutorials/topics/post-processing/count-objects-fomo.md) ### Build a keyword spotting or voice command model > I want to recognize spoken keywords or voice commands that run entirely on-device, including how to gather or synthesize the audio data. - [Follow an end-to-end keyword spotting project.](https://docs.edgeimpulse.com/tutorials/end-to-end/keyword-spotting.md) - [Train a keyword model with audio transfer learning.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/transfer-learning-keyword.md) - [Synthesize keyword audio when you lack real recordings.](https://docs.edgeimpulse.com/tutorials/topics/data/generate-keyword-data-google-tts.md) - [Reuse a ready-made keyword spotting dataset to get started.](https://docs.edgeimpulse.com/datasets/audio/keyword-spotting.md) ### Recognize motion or gestures from accelerometer data > I want to classify movements or gestures from IMU / accelerometer time-series data and deploy the model to a wearable or microcontroller. - [Follow an end-to-end motion / gesture recognition project.](https://docs.edgeimpulse.com/tutorials/end-to-end/motion-recognition.md) - [Extract frequency features from motion signals with spectral analysis.](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/spectral-analysis.md) - [Reuse a continuous motion dataset to prototype quickly.](https://docs.edgeimpulse.com/datasets/time-series/continuous-motion-recognition.md) ### Detect anomalies or defects (visual or sensor) > I want to detect defects, faults, or unusual patterns by learning what 'normal' looks like, using visual anomaly detection or sensor-based anomaly detection. - [Follow an end-to-end visual anomaly / defect detection project.](https://docs.edgeimpulse.com/tutorials/end-to-end/visual-anomaly-detection.md) - [Detect visual defects with FOMO-AD when defect samples are scarce.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/visual-anomaly-detection-fomo-ad.md) - [Flag abnormal sensor readings with K-means anomaly detection.](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/anomaly-detection-k-means.md) ### Upload or collect data programmatically via the API and SDK > I want to push samples into a project or pull data out programmatically using the Ingestion API or the Python SDK, including structured labels. - [Reference the Ingestion API for uploading samples and device data.](https://docs.edgeimpulse.com/apis/ingestion.md) - [Upload samples with structured multi-label metadata.](https://docs.edgeimpulse.com/tutorials/tools/apis/ingestion/ingest-structured-label-data.md) - [Use the Python SDK to move data in and out of projects.](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python.md) - [Upload and download project data from Python step by step.](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/upload-download-data.md) ### Optimize and fit a model to device constraints > My model needs to fit a device's memory and latency budget - I want to search architectures, compare experiments, profile on-device cost, and validate accuracy. - [Search for models and DSP configurations that meet device constraints.](https://docs.edgeimpulse.com/studio/projects/eon-tuner.md) - [Compare candidate impulses on accuracy, latency, and memory.](https://docs.edgeimpulse.com/studio/projects/experiments.md) - [Validate the chosen model on held-out data before shipping.](https://docs.edgeimpulse.com/studio/projects/model-testing.md) ### Choose a deployment target and runtime > I have a trained impulse and need to decide how to run it - MCU C++ library, Linux, browser/WebAssembly, or a container - and which integration guide to follow. - [Browse every deployment option and pick the runtime for your target.](https://docs.edgeimpulse.com/hardware/deployments.md) - [Run inference from the portable C++ library on microcontrollers.](https://docs.edgeimpulse.com/hardware/deployments/run-cpp.md) - [Run inference on Linux devices with the EIM runtime.](https://docs.edgeimpulse.com/hardware/deployments/run-linux-eim.md) - [Run inference in the browser with WebAssembly.](https://docs.edgeimpulse.com/hardware/deployments/run-webassembly-browser.md) - [Run inference in a container for servers or edge gateways.](https://docs.edgeimpulse.com/hardware/deployments/run-docker.md) ### Choose the right API or interface > I need to decide which programmatic interface to use - Studio API, Ingestion API, Remote Management API, the Python SDK, or the CLI. - [Automate projects, training, jobs, deployments, and organizations.](https://docs.edgeimpulse.com/apis/studio.md) - [Upload samples and device data into a project.](https://docs.edgeimpulse.com/apis/ingestion.md) - [Connect and manage device communication workflows.](https://docs.edgeimpulse.com/apis/remote-management.md) - [Use the supported Python interface for Studio workflows.](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python.md) - [Connect devices, upload data, run impulses, and build custom blocks from the CLI.](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli.md) ### Automate training and deployment (MLOps / CI/CD) > I want to automate model builds, testing, and deployment with the Studio API and CI pipelines like GitHub Actions for continuous edge MLOps. - [Drive builds, tests, and deployments from the Studio API.](https://docs.edgeimpulse.com/apis/studio.md) - [Wire training and deployment into a GitHub Actions pipeline.](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/deploy-github-actions.md) - [Version, deploy, monitor, and update edge ML systems.](https://docs.edgeimpulse.com/knowledge/concepts/lifecycle/lifecycle-management.md) ### Ship over-the-air (OTA) model updates to deployed devices > I want to update the model on already-deployed devices over the air, and understand the OTA update lifecycle and firmware integration. - [Understand the over-the-air model update lifecycle.](https://docs.edgeimpulse.com/knowledge/concepts/lifecycle/ota-model-updates.md) - [Implement OTA model updates in a C++ firmware application.](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-cpp.md) - [Implement OTA model updates on Arduino devices.](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-arduino.md) ### Import and deploy my own pretrained model (BYOM) > I already have a trained model and want to bring it into Edge Impulse, including custom or freeform input/output formats, and deploy it. - [Bring your own model into a project and profile it for a target.](https://docs.edgeimpulse.com/studio/projects/dashboard/byom.md) - [Deploy a model with custom or freeform input/output formats.](https://docs.edgeimpulse.com/tutorials/topics/machine-learning/deploy-freeform-model.md) ### Generate synthetic data when I lack real samples > I don't have enough real-world data and want to generate synthetic images, audio, or time-series data to train my model. - [Generate synthetic training data from inside Studio.](https://docs.edgeimpulse.com/studio/projects/data-acquisition/synthetic-data.md) - [Understand when and how synthetic data helps model training.](https://docs.edgeimpulse.com/knowledge/concepts/data-engineering/synthetic-data.md) - [Create synthetic image data with a generative model.](https://docs.edgeimpulse.com/tutorials/topics/data/generate-image-data-dall-e.md) ### Give an AI coding agent knowledge of Edge Impulse > I want my AI coding agent to understand the Edge Impulse Studio and Ingestion APIs and my target hardware using Agent Skills. - [Package Edge Impulse knowledge as an Agent Skill.](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/create-edge-impulse-skill.md) - [Build a companion skill scoped to your own project or hardware.](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/create-a-companion-skill.md) - [Reuse tested prompts for common Edge Impulse agent tasks.](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/prompt-library.md) ### Troubleshoot when official docs and community answers disagree > I hit an error or unexpected behavior and want to check the FAQ and glossary first, then community answers - preferring current official documentation when sources conflict. - [Check answers to common questions and known issues first.](https://docs.edgeimpulse.com/knowledge/faq.md) - [Clarify unfamiliar Edge Impulse terminology.](https://docs.edgeimpulse.com/knowledge/glossary.md) - [Search community answers, treating them as user-generated and preferring official docs when they conflict.](https://forum.edgeimpulse.com) ## Knowledge - [Edge Impulse Documentation](https://docs.edgeimpulse.com/index.md): Find guides, tutorials, API references, and hardware documentation for building edge AI with Edge Impulse. - [Knowledge](https://docs.edgeimpulse.com/knowledge.md): Explore edge AI guides, concepts, metrics, and courses for learning Edge Impulse workflows. - [FAQ](https://docs.edgeimpulse.com/knowledge/faq.md): Find answers to common Edge Impulse questions about data, processing, learning, deployment, sharing, and accounts. - [Glossary](https://docs.edgeimpulse.com/knowledge/glossary.md): Look up common Edge Impulse, edge AI, embedded systems, and machine learning terms. - [Getting started for beginners](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-beginners.md): Get started with Edge Impulse by creating a project, collecting data, training, deploying, and exploring tutorials. - [Getting started for embedded engineers](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-embedded-engineers.md): Get started with Edge Impulse as an embedded engineer collecting data, training models, and deploying to hardware. - [Getting started for machine learning practitioners](https://docs.edgeimpulse.com/knowledge/guides/getting-started-for-ml-practitioners.md): Get started with Edge Impulse as an ML practitioner using existing datasets, models, and deployment workflows. - [Combining impulses](https://docs.edgeimpulse.com/knowledge/guides/combining-impulses.md): Compare multi-impulse, multi-model, and sensor fusion approaches for combining Edge Impulse project outputs. - [Increasing model performance](https://docs.edgeimpulse.com/knowledge/guides/increasing-model-performance.md): Diagnose poor impulse performance and improve data balance, training, architecture, and quantized model results. - [Lowering compute time](https://docs.edgeimpulse.com/knowledge/guides/lowering-compute-time.md): Reduce Edge Impulse training compute time by shrinking datasets, changing epochs, adjusting batches, and simplifying models. - [Health reference design](https://docs.edgeimpulse.com/knowledge/guides/reference-designs/health-reference-design.md): Build a wearable health reference workflow that synchronizes, validates, queries, transforms, and pipelines clinical data. - [Synchronizing clinical data with a bucket](https://docs.edgeimpulse.com/knowledge/guides/reference-designs/health-reference-design/synchronizing-clinical-data.md): Synchronize PPG-DaLiA clinical study files into organizational datasets and prepare them for validation and transformation. - [Validating clinical data](https://docs.edgeimpulse.com/knowledge/guides/reference-designs/health-reference-design/validating-clinical-data.md): Use dataset checklists and metadata to validate clinical study files before transforming and training models. - [Querying clinical data](https://docs.edgeimpulse.com/knowledge/guides/reference-designs/health-reference-design/querying-clinical-data.md): Use organizational dataset filters to query clinical data items, files, metadata, and shared results. - [Transforming clinical data](https://docs.edgeimpulse.com/knowledge/guides/reference-designs/health-reference-design/transforming-clinical-data.md): Create a Python transformation block that converts PPG-DaLiA clinical data into machine learning features. - [Audio feature extraction](https://docs.edgeimpulse.com/knowledge/concepts/data-engineering/audio-feature-extraction.md): Understand audio feature extraction techniques in Edge Impulse, including spectrogram, MFCC, and MFE blocks. - [Motion feature extraction](https://docs.edgeimpulse.com/knowledge/concepts/data-engineering/motion-feature-extraction.md): Learn how spectral features extract motion signals for activity, gesture, and vibration models in Edge Impulse. - [Synthetic data](https://docs.edgeimpulse.com/knowledge/concepts/data-engineering/synthetic-data.md): Generate synthetic datasets in Edge Impulse and find tutorials for image, audio, and simulation workflows. - [Data augmentation](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/data-augmentation.md): Improve model generalization by applying data augmentation during training and comparing results against baseline models. - [Neural networks](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks.md): Understand neural network concepts, layers, training, and optimization for edge AI applications in Edge Impulse. - [Layers](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/layers.md): Explore common neural network layer types and how to combine them in Edge Impulse expert mode. - [Activation functions](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/activation-functions.md): Choose and customize neural network activation functions for hidden and output layers in Edge Impulse models. - [Loss functions](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/loss-functions.md): Choose and customize loss functions for regression and classification neural networks in Edge Impulse expert mode. - [Optimizers](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/optimizers.md): Understand optimizer choices and how to customize Keras optimizers for Edge Impulse neural network training. - [Learned optimizer (VeLO)](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/learned-optimizer-velo.md): Enable and evaluate the VeLO learned optimizer for training Edge Impulse neural network models. - [Epochs](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/neural-networks/epochs.md): Understand training epochs, batch size, overfitting, and early stopping when training neural networks in Edge Impulse. - [On-device learning](https://docs.edgeimpulse.com/knowledge/concepts/machine-learning/on-device-learning.md): Review on-device learning scenarios and Edge Impulse options for training or adapting models near production devices. - [Lifecycle management](https://docs.edgeimpulse.com/knowledge/concepts/lifecycle/lifecycle-management.md): Plan lifecycle management for Edge Impulse projects, including MLOps concepts, OTA infrastructure, and continuous learning loops. - [OTA model updates](https://docs.edgeimpulse.com/knowledge/concepts/lifecycle/ota-model-updates.md): Learn the OTA model update flow for detecting, downloading, validating, and deploying updated Edge Impulse impulses. - [What is embedded ML, anyway?](https://docs.edgeimpulse.com/knowledge/concepts/what-is-embedded-machine-learning-anyway.md): Understand embedded machine learning, when it helps, and why on-device inference benefits connected products. - [What is edge machine learning (edge ML)?](https://docs.edgeimpulse.com/knowledge/concepts/what-is-edge-machine-learning.md): Learn how edge machine learning moves inference from cloud servers to local devices and embedded systems. - [Definitions](https://docs.edgeimpulse.com/knowledge/metrics/definitions.md): Review common model evaluation metrics and formulas for classification, regression, and object detection tasks. - [Inference performance](https://docs.edgeimpulse.com/knowledge/metrics/inference-performance.md): Compare typical Edge Impulse model latency, RAM, and ROM usage across supported microcontroller classes. - [Model evaluation](https://docs.edgeimpulse.com/knowledge/metrics/model-evaluation.md): Choose evaluation metrics for classification, regression, and object detection models deployed on edge devices. - [Edge AI Fundamentals](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals.md): Start the Edge AI Fundamentals course and follow its lessons, quizzes, and certificate path. - [Introduction to edge AI](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/intro-to-edge-ai.md): Start the Edge AI Fundamentals course with an overview of edge AI concepts, lessons, quizzes, and certification. - [What is edge computing?](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/what-is-edge-computing.md): Learn how edge computing differs from cloud computing and why local processing can reduce latency and bandwidth. - [What is machine learning (ML)?](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/what-is-machine-learning.md): Understand machine learning, its relationship to AI, its history, model categories, and training workflows. - [What is edge AI?](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/what-is-edge-ai.md): Understand edge AI training, inference, deployment benefits, limitations, examples, and market trends. - [How to choose an edge AI device](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/how-to-choose-an-edge-ai-device.md): Choose edge AI hardware by comparing use cases, design constraints, and device processor classes. - [Edge AI lifecycle](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/edge-ai-lifecycle.md): Follow the edge AI lifecycle from project scoping through data collection, model development, deployment, and maintenance. - [What is edge MLOps?](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/what-is-edge-mlops.md): Learn edge MLOps principles for automating data, model, deployment, monitoring, and maintenance workflows. - [What is Edge Impulse?](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/what-is-edge-impulse.md): See how Edge Impulse supports data collection, model training, deployment, and edge MLOps workflows. - [Case study: Izoelektro smart grid monitoring](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/case-study-izoelektro.md): Study how Izoelektro uses edge AI anomaly detection to monitor remote power grid equipment. - [Test and certification](https://docs.edgeimpulse.com/knowledge/courses/edge-ai-fundamentals/test-and-certification.md): Finish the Edge AI Fundamentals course with further learning resources, the comprehensive test, and certification details. - [Introduction to Embedded Machine Learning](https://docs.edgeimpulse.com/knowledge/courses/introduction-embedded-ml.md): Open the Coursera course on introductory embedded machine learning with Edge Impulse. - [Computer Vision with Embedded Machine Learning](https://docs.edgeimpulse.com/knowledge/courses/computer-vision-embedded-ml.md): Open the Coursera course on computer vision with embedded machine learning. ## Studio - [Studio](https://docs.edgeimpulse.com/studio.md): Build, train, and deploy edge AI models in Edge Impulse Studio using projects and organizations. - [Account settings](https://docs.edgeimpulse.com/studio/account-settings.md): Manage your Edge Impulse account profile, password, plan, multi-factor authentication, secrets, and organization access. - [Dashboard](https://docs.edgeimpulse.com/studio/projects/dashboard.md): Manage project settings, visibility, collaborators, README content, performance options, and API credentials. - [API keys](https://docs.edgeimpulse.com/studio/projects/dashboard/api-keys.md): Create project API keys and HMAC keys for programmatic access, development tooling, and signed ingestion. - [Target device](https://docs.edgeimpulse.com/studio/projects/dashboard/target-device.md): Set target hardware constraints so Edge Impulse Studio can recommend defaults and budget model performance for deployment. - [Bring your own model (BYOM)](https://docs.edgeimpulse.com/studio/projects/dashboard/byom.md): Upload pretrained models, inspect their behavior, and optionally add Edge Impulse DSP preprocessing blocks. - [Devices](https://docs.edgeimpulse.com/studio/projects/devices.md): Connect devices to collect data, run live classification, and deploy pre-built firmware from your project. - [Data acquisition](https://docs.edgeimpulse.com/studio/projects/data-acquisition.md): Explore the project data acquisition tabs for datasets, sources, synthetic data, labeling, and CSV imports. - [Dataset](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset.md): Review collected samples, dataset splits, filters, previews, and sample editing options in the Dataset tab. - [Uploader](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/uploader.md): Upload existing samples and datasets through Edge Impulse Studio with supported labels and annotation formats. - [Define dataset split](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/splits.md): Define training, validation, and test splits with ratios, stratification, and metadata-based grouping controls. - [Metadata](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/metadata.md): Use metadata keys to organize samples, analyze distributions, and control training and validation splits. - [Multi-label data](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/multi-label.md): Use multi-label samples to preserve long time-series context and choose labeling strategies for model windows. - [Tabular data](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/tabular-data.md): Import pre-processed CSV data as tabular samples and test tabular models in Edge Impulse Studio. - [Advanced dataset settings](https://docs.edgeimpulse.com/studio/projects/data-acquisition/dataset/advanced-settings.md): Configure explicit validation sets, disable reset behavior, and adjust dataset settings from the Advanced settings modal. - [Data explorer](https://docs.edgeimpulse.com/studio/projects/data-acquisition/data-explorer.md): Visualize dataset embeddings to inspect clusters, identify outliers, and relabel samples in Edge Impulse Studio. - [Data sources](https://docs.edgeimpulse.com/studio/projects/data-acquisition/data-sources.md): Connect external storage to a project and automate ingestion with labeling, transformations, schedules, and webhooks. - [Synthetic data](https://docs.edgeimpulse.com/studio/projects/data-acquisition/synthetic-data.md): Generate image, audio, and time-series synthetic data through built-in blocks or custom integrations. - [Labeling queue](https://docs.edgeimpulse.com/studio/projects/data-acquisition/labeling-queue.md): Use the Labeling queue to create bounding boxes and get AI-assisted suggestions for object detection datasets. - [AI labeling](https://docs.edgeimpulse.com/studio/projects/data-acquisition/ai-labeling.md): Configure AI labeling actions that use built-in or custom blocks to label project data automatically. - [CSV Wizard](https://docs.edgeimpulse.com/studio/projects/data-acquisition/csv-wizard.md): Import custom CSV files by mapping metadata, sensor values, labels, and sample windows. - [Auto-labeler](https://docs.edgeimpulse.com/studio/projects/data-acquisition/auto-labeler.md): Use embeddings to group unlabeled object detection images and assign bounding boxes more quickly. - [Experiments](https://docs.edgeimpulse.com/studio/projects/experiments.md): Compare multiple impulses in one project and export accuracy and performance results. - [EON Tuner](https://docs.edgeimpulse.com/studio/projects/eon-tuner.md): Run hardware-aware searches to compare impulses and select efficient configurations for your target device. - [Search space](https://docs.edgeimpulse.com/studio/projects/eon-tuner/search-space.md): Constrain EON Tuner searches with templates that define input, processing, and learning block combinations. - [Impulse design](https://docs.edgeimpulse.com/studio/projects/impulse-design.md): Build impulses by combining input, processing, and learning blocks for different project types. - [Processing blocks](https://docs.edgeimpulse.com/studio/projects/processing-blocks.md): Extract features with built-in processing blocks, custom DSP code, and feature normalization options. - [Feature explorer](https://docs.edgeimpulse.com/studio/projects/processing-blocks/feature-explorer.md): Use feature plots to inspect class separation, outliers, and dimensionality reduction results. - [Audio MFCC](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/audio-mfcc.md): Configure MFCC audio feature extraction parameters for speech recognition and related sound models. - [Audio MFE](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/audio-mfe.md): Extract Mel-filterbank energy features from audio signals and tune the parameters used during feature generation. - [Audio Syntiant](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/audio-syntiant.md): Configure Syntiant-specific audio feature extraction for projects that target NDP101 or NDP120 chips. - [EEG](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/eeg.md): Preprocess EEG time-series data with filtering, motion artifact removal, scaling, and spectral feature extraction. - [Flatten](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/flatten.md): Summarize slow-moving sensor windows with statistical features such as averages, extrema, RMS, and variance. - [HR/HRV](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/hr-hrv.md): Extract heart rate and heart rate variability metrics from PPG or ECG signals for edge deployments. - [Image](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/image.md): Normalize image inputs and choose color depth settings for computer vision impulses. - [IMU Syntiant](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/imu-syntiant.md): Rescale IMU input data for Syntiant NDP101 and NDP120 chip requirements. - [Raw data](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/raw-data.md): Pass preprocessed samples through the impulse with optional feature normalization and axis scaling. - [Spectral analysis](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/spectral-analysis.md): Extract frequency-domain and statistical features from sensor signals using FFT or wavelet analysis. - [Spectrogram](https://docs.edgeimpulse.com/studio/projects/processing-blocks/blocks/spectrogram.md): Convert continuous signals into time-frequency features for audio and sensor classification use cases. - [Custom processing blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-processing-blocks.md): Build processing blocks that transform raw signals into features and visualizations for impulse design. - [Learning blocks](https://docs.edgeimpulse.com/studio/projects/learning-blocks.md): Train models with built-in learning blocks, configure neural network settings, and access expert mode. - [Expert mode](https://docs.edgeimpulse.com/studio/projects/learning-blocks/expert-mode.md): Customize Keras learning blocks with expert mode and edit architectures beyond the visual editor. - [Training graphs](https://docs.edgeimpulse.com/studio/projects/learning-blocks/training-graphs.md): Inspect accuracy and loss graphs, open TensorBoard, and compare training runs across experiments. - [Anomaly detection](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/anomaly-detection-custom.md): Host custom anomaly detection algorithms in an impulse and understand the expected learning block workflow. - [Anomaly detection (GMM)](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/anomaly-detection-gmm.md): Train and test GMM anomaly detection models with selected features, components, and confidence thresholds. - [Anomaly detection (K-means)](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/anomaly-detection-k-means.md): Train K-means anomaly detection models and choose features for detecting outliers in sensor data. - [Classification](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/classification.md): Train neural network classifiers and review model performance after feature extraction. - [Classical ML](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/classical-ml.md): Train scikit-learn based classical ML classifiers and regressors from Edge Impulse processing block features. - [Object detection](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection.md): Compare built-in object detection architectures and choose a model for bounding-box projects. - [YOLO-Pro](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection/yolo-pro.md): Use YOLO-Pro object detection architectures in an impulse and configure training settings. - [FOMO](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection/fomo.md): Use FOMO to detect object centroids on constrained devices and customize the model in expert mode. - [MobileNetV2 SSD FPN](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/object-detection/mobilenetv2-ssd-fpn.md): Configure MobileNetV2 SSD FPN-Lite for bounding-box object detection with transfer learning. - [Regression](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/regression.md): Train regression models from numeric labels and test continuous predictions on unseen samples. - [Transfer learning (images)](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/transfer-learning-images.md): Use image transfer learning with pre-trained networks and expert mode for classification projects. - [Transfer learning (keyword spotting)](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/transfer-learning-keyword.md): Fine-tune pre-trained keyword spotting models and adjust neural network settings for audio projects. - [Visual anomaly detection (FOMO-AD)](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/visual-anomaly-detection-fomo-ad.md): Train visual anomaly detection models with PatchCore or GMM scoring and set confidence thresholds. - [Custom learning blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-learning-blocks.md): Create learning blocks that train custom models from project data and return deployable model artifacts. - [NVIDIA TAO](https://docs.edgeimpulse.com/studio/projects/learning-blocks/blocks/nvidia-tao.md): Understand the deprecation of NVIDIA TAO models and choose supported Edge Impulse alternatives. - [Retrain model](https://docs.edgeimpulse.com/studio/projects/retrain-model.md): Regenerate features and retrain an existing impulse after adding data to a project. - [Live classification](https://docs.edgeimpulse.com/studio/projects/live-classification.md): Validate trained models with live data from devices, phones, computers, or uploaded samples. - [Model testing](https://docs.edgeimpulse.com/studio/projects/model-testing.md): Evaluate trained models on test data, inspect confusion matrices, and set confidence thresholds. - [Performance calibration](https://docs.edgeimpulse.com/studio/projects/performance-calibration.md): Tune event detection post-processing with streaming tests and compare recommended configurations. - [Object tracking](https://docs.edgeimpulse.com/studio/projects/post-processing/object-tracking.md): Track object detections across frames, configure trace settings, and reduce jitter in post-processing. - [Deployment](https://docs.edgeimpulse.com/studio/projects/deployment.md): Choose deployment targets, download builds, and compare model optimization options for edge devices. - [EON Compiler](https://docs.edgeimpulse.com/studio/projects/deployment/eon-compiler.md): Compile models into optimized C++ and compare EON Compiler resource savings across supported architectures. - [EON Operator Matrix (TensorFlow 2.19)](https://docs.edgeimpulse.com/studio/projects/deployment/eon-compiler-operator-matrix-2-19.md): View the EON Compiler operator support matrix in a full-width layout for TensorFlow 2.19. - [Custom deployment blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-deployment-blocks.md): Create deployment blocks that package trained models into custom artifacts for devices, services, or toolchains. - [Multi-project deployment](https://docs.edgeimpulse.com/studio/organizations/multi-project-deployment.md): Bundle multiple Edge Impulse projects and impulses into one deployment package for coordinated on-device inference. - [Versioning](https://docs.edgeimpulse.com/studio/projects/versioning.md): Create, view, and revert project versions to track machine learning workflow changes over time. - [Dashboard](https://docs.edgeimpulse.com/studio/organizations/dashboard.md): Monitor organization usage, data, projects, campaigns, blocks, and members from the organization dashboard. - [Data campaigns](https://docs.edgeimpulse.com/studio/organizations/data-campaigns.md): Create organization data campaigns to collect files through upload portals and route them into datasets. - [Users](https://docs.edgeimpulse.com/studio/organizations/users.md): Understand organization roles and compare the permissions available to Admins, Members, and Guests. - [Projects](https://docs.edgeimpulse.com/studio/organizations/projects.md): View and manage the Edge Impulse projects that belong to your organization. - [Data](https://docs.edgeimpulse.com/studio/organizations/data.md): Centralize organization datasets, connect storage buckets, and prepare files for importing into Edge Impulse projects. - [Cloud data storage](https://docs.edgeimpulse.com/studio/organizations/data/cloud-data-storage.md): Connect organization or project data sources to S3, Google Cloud Storage, Azure Blob Storage, or S3-compatible buckets. - [Data pipelines](https://docs.edgeimpulse.com/studio/organizations/data-pipelines.md): Combine transformation blocks into scheduled pipelines that process organization data and trigger notifications or webhooks. - [Data transformation](https://docs.edgeimpulse.com/studio/organizations/data-transformation.md): Run transformation jobs on organization datasets to generate files, update metadata, or import data into projects. - [Transformation blocks](https://docs.edgeimpulse.com/studio/organizations/transformation-blocks.md): Run pre-built or custom transformation blocks to process files, folders, or standalone organization jobs. - [Time-series data augmentation](https://docs.edgeimpulse.com/studio/organizations/transformation-blocks/blocks/time-series-data-augmentation.md): Use the time-series data augmentation block to generate synthetic sensor signals for organization datasets. - [Upload portals](https://docs.edgeimpulse.com/studio/organizations/upload-portals.md): Configure upload portals that let external contributors send files into organization storage buckets. - [Custom blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks.md): Create and manage organization custom blocks for labeling, processing, learning, transformation, deployment, and synthetic data workflows. - [Custom AI labeling blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-ai-labeling-blocks.md): Build custom AI labeling blocks that generate labels or bounding boxes from data stored in Edge Impulse. - [Custom synthetic data blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-synthetic-data-blocks.md): Create synthetic data blocks that generate samples and upload them to Edge Impulse through the Ingestion API. - [Custom transformation blocks](https://docs.edgeimpulse.com/studio/organizations/custom-blocks/custom-transformation-blocks.md): Create transformation blocks that process organization data, update metadata, and run from jobs or data sources. - [Secrets](https://docs.edgeimpulse.com/studio/organizations/secrets.md): Store organization-level secret values that custom blocks and automation can reference securely. ## Hardware - [Hardware](https://docs.edgeimpulse.com/hardware.md): Find supported edge AI devices, boards, and deployment options for running Edge Impulse models. - [Linux Process Overview](https://docs.edgeimpulse.com/hardware/porting/linux-inference/linux-inference-process.md): Validate that a Linux device can run Edge Impulse models before deeper hardware integration. - [Inferencing with Edge Impulse Linux CLI](https://docs.edgeimpulse.com/hardware/porting/linux-inference/linux-inference-cli.md): Use the Edge Impulse Linux CLI to download and run model inference on Linux targets. - [Inferencing with Edge Impulse Linux C++ SDK](https://docs.edgeimpulse.com/hardware/porting/linux-inference/linux-inference-cpp.md): Build C++ inference applications for Linux targets using exported Edge Impulse model libraries. - [Inferencing with Edge Impulse Linux Python SDK](https://docs.edgeimpulse.com/hardware/porting/linux-inference/linux-inference-python.md): Run Edge Impulse Linux EIM models from Python applications on supported Linux targets. - [Advantech ICAM-540](https://docs.edgeimpulse.com/hardware/devices/advantech-icam-540.md): Connect an Advantech ICAM-540 to Edge Impulse, configure its camera sensor, and deploy accelerated image models. - [BrickML](https://docs.edgeimpulse.com/hardware/devices/brickml.md): Connect the BrickML industrial reference device to Edge Impulse and build predictive maintenance models. - [CODICO Triple Vision Industrial AI Camera](https://docs.edgeimpulse.com/hardware/devices/jmo-triple-vision-camera.md): Set up the CODICO Triple Vision camera, connect it to Edge Impulse, and deploy accelerated vision models. - [OnLogic FR101](https://docs.edgeimpulse.com/hardware/devices/onlogic-fr101.md): Set up the OnLogic FR101 for Edge Impulse Linux inference with Qualcomm QNN acceleration. - [OnLogic Helix 521](https://docs.edgeimpulse.com/hardware/devices/onlogic-hx521.md): Set up the OnLogic Helix 521 as a Linux target for Edge Impulse data collection and inference. - [OnLogic ML100G](https://docs.edgeimpulse.com/hardware/devices/onlogic-ml100g.md): Set up the OnLogic ML100G as a Linux target for Edge Impulse data collection and inference. - [Seeed reComputer Jetson](https://docs.edgeimpulse.com/hardware/devices/seeed-recomputer-jetson.md): Connect a Seeed reComputer Jetson to Edge Impulse and deploy models back to the device. - [Seeed SenseCAP A1101](https://docs.edgeimpulse.com/hardware/devices/seeed-sensecap-a1101.md): Update SenseCAP A1101 firmware, connect the camera to Edge Impulse, and deploy vision models. - [macOS devices](https://docs.edgeimpulse.com/hardware/devices/macos.md): Use a Mac as an Edge Impulse Linux target for data collection and local model inference. - [Linux x86_64 devices](https://docs.edgeimpulse.com/hardware/devices/linux-x86_64.md): Set up Linux x86_64 devices to collect data and run Edge Impulse models locally. - [Mobile phone](https://docs.edgeimpulse.com/hardware/devices/mobile-phone.md): Connect a mobile phone to Edge Impulse for browser-based data collection and local WebAssembly inference. - [Ambiq Apollo4 evaluation boards](https://docs.edgeimpulse.com/hardware/boards/ambiq-apollo4.md): Set up Ambiq Apollo4 evaluation boards for audio and image data collection, training, testing, and deployment. - [Arducam Pico4ML TinyML Dev Kit](https://docs.edgeimpulse.com/hardware/boards/arducam-pico4ml-tinyml-dev-kit.md): Connect the Arducam Pico4ML TinyML Dev Kit to Edge Impulse and deploy trained models. - [Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/hardware/boards/arduino-nano-33-ble-sense.md): Set up the Arduino Nano 33 BLE Sense, collect sensor or camera data, and deploy trained models. - [Arduino Nicla Sense ME](https://docs.edgeimpulse.com/hardware/boards/arduino-nicla-sense-me.md): Connect the Arduino Nicla Sense ME to Edge Impulse, collect sensor data, and deploy Arduino examples. - [Arduino Nicla Vision](https://docs.edgeimpulse.com/hardware/boards/arduino-nicla-vision.md): Set up the Arduino Nicla Vision with official firmware or ingestion sketches for edge AI projects. - [Arduino Portenta H7](https://docs.edgeimpulse.com/hardware/boards/arduino-portenta-h7.md): Connect the Arduino Portenta H7 and vision shield to Edge Impulse and deploy trained models. - [Blues Wireless Swan](https://docs.edgeimpulse.com/hardware/boards/blues-wireless-swan.md): Use the Blues Wireless Swan community tutorial to collect accelerometer data and deploy Edge Impulse models. - [Espressif ESP-EYE](https://docs.edgeimpulse.com/hardware/boards/espressif-esp32.md): Flash the Espressif ESP-EYE firmware, connect it to Edge Impulse, and deploy models for ESP32 boards. - [Himax WE-I Plus](https://docs.edgeimpulse.com/hardware/boards/himax-we-i-plus.md): Flash the Himax WE-I Plus firmware, connect the board to Edge Impulse, and troubleshoot common setup issues. - [Infineon CY8CKIT-062-BLE Pioneer Kit](https://docs.edgeimpulse.com/hardware/boards/infineon-cy8ckit-062-ble.md): Flash the CY8CKIT-062-BLE Pioneer Kit base firmware and connect it to Edge Impulse for sensor projects. - [Infineon CY8CKIT-062S2 Pioneer Kit](https://docs.edgeimpulse.com/hardware/boards/infineon-cy8ckit-062s2.md): Flash the CY8CKIT-062S2 Pioneer Kit base firmware and connect it to Edge Impulse for sensor projects. - [Nordic Semi nRF52840 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf52840-dk.md): Flash the nRF52840 DK firmware, attach the X-NUCLEO-IKS02A1 shield, and connect to Edge Impulse. - [Nordic Semi nRF5340 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf5340-dk.md): Flash the nRF5340 DK firmware, attach the X-NUCLEO-IKS02A1 shield, and connect to Edge Impulse. - [Nordic Semi nRF54L15 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf54L15-dk.md): Wire the nRF54L15 DK to an IKS02A1 shield and connect it to Edge Impulse for sensor projects. - [Nordic Semi nRF9160 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf9160-dk.md): Flash the nRF9160 DK firmware, attach the X-NUCLEO-IKS02A1 shield, and connect to Edge Impulse. - [Nordic Semi nRF9161 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf9161-dk.md): Set up the Nordic nRF9161 DK with Edge Impulse firmware and connect supported sensors for data collection. - [Nordic Semi nRF9151 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf9151-dk.md): Flash the nRF9151 DK firmware, attach the X-NUCLEO-IKS02A1 shield, and connect to Edge Impulse. - [Nordic Semi nRF7002 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf7002-dk.md): Connect the nRF7002 DK to Edge Impulse over Wi-Fi and flash motion recognition firmware. - [Nordic Semi Thingy:53](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-thingy53.md): Set up Nordic Thingy:53 with Edge Impulse, connect by serial or mobile app, and use WiFi expansion. - [Nordic Semi Thingy:91](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-thingy91.md): Update Nordic Thingy:91 firmware, connect it to Edge Impulse, and start building sensor models. - [Open MV Cam H7 Plus](https://docs.edgeimpulse.com/hardware/boards/openmv-cam-h7-plus.md): Connect the OpenMV Cam H7 Plus to Edge Impulse and find image collection and deployment tutorials. - [Particle Boron](https://docs.edgeimpulse.com/hardware/boards/particle-boron.md): Wire Particle Boron sensors, install Particle tools, and flash firmware for Edge Impulse workflows. - [Particle Photon 2](https://docs.edgeimpulse.com/hardware/boards/particle-photon-2.md): Wire Particle Photon 2 sensors, flash firmware, collect data, and configure deployment or webhook workflows. - [Qualcomm QCC744M Evaluation Kit](https://docs.edgeimpulse.com/hardware/boards/qualcomm-qcc744m_evk.md) - [RAKwireless WisBlock](https://docs.edgeimpulse.com/hardware/boards/rakwireless-wisblock.md): Use RAKwireless WisBlock cores with Edge Impulse guides to collect data and run local inference. - [Raspberry Pi Pico](https://docs.edgeimpulse.com/hardware/boards/raspberry-pi-pico.md): Set up Raspberry Pi Pico or RP2350 boards with Edge Impulse firmware and supported external sensors. - [Renesas CK-RA6M5 Cloud Kit](https://docs.edgeimpulse.com/hardware/boards/renesas-ck-ra6m5.md): Flash the Renesas CK-RA6M5 Cloud Kit, connect it to Edge Impulse, and collect sensor data. - [Renesas EK-RA8D1](https://docs.edgeimpulse.com/hardware/boards/renesas-ek-ra8d1.md): Flash the Renesas EK-RA8D1, connect it to Edge Impulse, and build image-based models. - [Seeed Wio Terminal](https://docs.edgeimpulse.com/hardware/boards/seeed-wio-terminal.md): Connect the Seeed Wio Terminal to Edge Impulse and deploy Arduino library impulses to the board. - [Seeed XIAO nRF52840 Sense](https://docs.edgeimpulse.com/hardware/boards/seeed-xiao-nrf52840-sense.md): Connect the Seeed XIAO nRF52840 Sense to Edge Impulse and deploy Arduino library models to the board. - [Seeed XIAO ESP32 S3 Sense](https://docs.edgeimpulse.com/hardware/boards/seeed-xiao-esp32s3-sense.md): Connect the Seeed XIAO ESP32S3 Sense to Edge Impulse and deploy Arduino library models to the board. - [SiLabs Thunderboard Sense 2](https://docs.edgeimpulse.com/hardware/boards/silabs-thunderboard-sense-2.md): Flash the Thunderboard Sense 2 firmware, connect the board to Edge Impulse, and use the BLE demo. - [Sony Spresense](https://docs.edgeimpulse.com/hardware/boards/sony-spresense.md): Set up Sony Spresense with supported add-ons, connect it to Edge Impulse, and use CommonSense sensor fusion. - [ST B-L475E-IOT01A](https://docs.edgeimpulse.com/hardware/boards/st-b-l475e-iot01a.md): Flash the ST B-L475E-IOT01A firmware, connect it to Edge Impulse, and troubleshoot WiFi setup issues. - [TI CC1352P Launchpad](https://docs.edgeimpulse.com/hardware/boards/ti-launchxl.md): Configure TI Launchpad booster packs, flash Edge Impulse firmware, and collect accelerometer or audio data. - [Alif Ensemble series kits](https://docs.edgeimpulse.com/hardware/boards/alif-ensemble.md): Set up Alif Ensemble kits, install required tooling, flash firmware, and connect the board to Edge Impulse. - [Ambiq Apollo5 evaluation boards](https://docs.edgeimpulse.com/hardware/boards/ambiq-apollo5.md): Set up Ambiq Apollo5 evaluation boards for audio and image data collection, training, testing, and deployment. - [Arduino Nicla Voice](https://docs.edgeimpulse.com/hardware/boards/arduino-nicla-voice.md): Set up the Arduino Nicla Voice for low-power audio and IMU model development with Edge Impulse. - [Avnet RASynBoard](https://docs.edgeimpulse.com/hardware/boards/avnet-rasynboard.md): Set up the Avnet RASynBoard firmware and connect the board to Edge Impulse for audio or IMU projects. - [Nordic Semi nRF54LM20 DK](https://docs.edgeimpulse.com/hardware/boards/nordic-semi-nrf54LM20-dk.md): Wire the nRF54LM20 DK to an IKS02A1 shield and deploy Edge Impulse models with Nordic Axon support. - [Himax WiseEye2 Module and ISM Devboard](https://docs.edgeimpulse.com/hardware/boards/himax-ism-wise-eye-2.md): Set up the Himax WiseEye2 ISM Devboard, flash Edge Impulse firmware, and run image inference. - [Seeed Grove - Vision AI Module](https://docs.edgeimpulse.com/hardware/boards/seeed-grove-vision-ai.md): Set up the Seeed Grove - Vision AI Module, flash firmware, connect it to Edge Impulse, and deploy models. - [Seeed Grove Vision AI Module V2 (WiseEye2)](https://docs.edgeimpulse.com/hardware/boards/seeed-grove-vision-ai-module-v2-wise-eye-2.md): Flash the Seeed Grove Vision AI Module V2, connect it to Edge Impulse, and deploy vision models. - [SiLabs xG24 Dev Kit](https://docs.edgeimpulse.com/hardware/boards/silabs-xg24-devkit.md): Flash the xG24 Dev Kit base firmware, connect it to Edge Impulse, and run the BLE demo. - [STMicroelectronics STM32N6570-DK](https://docs.edgeimpulse.com/hardware/boards/stm32n6570-dk.md): Program STM32N6570-DK firmware binaries, connect the kit to Edge Impulse, and deploy object detection models. - [Synaptics Katana EVK](https://docs.edgeimpulse.com/hardware/boards/synaptics-katana.md): Connect the Synaptics Katana EVK to Edge Impulse, flash firmware, and build board-specific machine learning projects. - [Syntiant Tiny ML Board](https://docs.edgeimpulse.com/hardware/boards/syntiant-tinyml-board.md): Flash audio or IMU firmware for the Syntiant TinyML Board and connect it to Edge Impulse. - [Arduino UNO Q](https://docs.edgeimpulse.com/hardware/boards/arduino-uno-q.md): Set up the Arduino UNO Q for Linux-based data collection, inference, and remote access with Edge Impulse. - [Microchip SAMA7G54](https://docs.edgeimpulse.com/hardware/boards/microchip-sama7.md): Prepare Microchip SAMA7G54 Linux images, connect the board to Edge Impulse, and run deployed models. - [Raspberry Pi 4](https://docs.edgeimpulse.com/hardware/boards/raspberry-pi-4.md): Set up Raspberry Pi 4 for Edge Impulse, connect peripherals, install the Linux CLI, and troubleshoot deployment. - [Raspberry Pi 5](https://docs.edgeimpulse.com/hardware/boards/raspberry-pi-5.md): Set up Raspberry Pi 5 with Raspberry Pi OS, install Edge Impulse tools, and troubleshoot deployment. - [Renesas RZ/G2L](https://docs.edgeimpulse.com/hardware/boards/renesas-rz-g2l.md): Build a Renesas RZ/G2L Yocto image, connect the board, and deploy supported Edge Impulse models. - [Texas Instruments SK-AM62](https://docs.edgeimpulse.com/hardware/boards/ti-sk-am62.md): Use the Texas Instruments SK-AM62 starter kit with Edge Impulse through TI setup instructions. - [Advantech AOM-2721 OSM](https://docs.edgeimpulse.com/hardware/boards/advantech-aom2721-osm.md): Set up the Advantech AOM-2721 OSM with Edge Impulse, collect data, profile models, and deploy. - [Arduino VENTUNO Q](https://docs.edgeimpulse.com/hardware/boards/arduino-ventuno-q.md): Set up the Arduino VENTUNO Q with Ubuntu, install the Edge Impulse Linux CLI, and run deployed models. - [AVNET RZBoard V2L](https://docs.edgeimpulse.com/hardware/boards/avenet-rz-v2l.md): Review the AVNET RZBoard V2L hardware and links for building an Edge Impulse-ready board image. - [BrainChip AKD1000](https://docs.edgeimpulse.com/hardware/boards/brainchip-akd1000.md): Set up BrainChip Akida hardware with Edge Impulse and deploy compatible models to AKD1000 targets. - [Digi ConnectCore 93 Development Kit](https://docs.edgeimpulse.com/hardware/boards/digi-ccimx93-dvk.md): Prepare the Digi ConnectCore 93 Development Kit, install Edge Impulse Linux, and run deployed impulses. - [IMDT RZ/V2H](https://docs.edgeimpulse.com/hardware/boards/imdt-rz-v2h.md): Build an IMDT RZ/V2H Yocto image with Edge Impulse support and deploy DRP-AI accelerated models. - [Innodisk EXEC-Q911](https://docs.edgeimpulse.com/hardware/boards/innodisk-exec-q911.md): Set up Ubuntu on the Innodisk EXEC-Q911, install Qualcomm dependencies, and deploy accelerated Edge Impulse models. - [MemryX MX3](https://docs.edgeimpulse.com/hardware/boards/memryx-mx3.md): Install MemryX MX3 dependencies, connect Edge Impulse to Linux targets, and deploy MX3-accelerated models. - [MistyWest MistySOM RZ/V2L](https://docs.edgeimpulse.com/hardware/boards/mistywest-rz-v2l.md): Review MistySOM RZ/V2L hardware capabilities and support links for using the board with Edge Impulse. - [NXP i.MX 8M Plus EVK](https://docs.edgeimpulse.com/hardware/boards/nxp-imx8-evk.md): Set up the NXP i.MX 8M Plus EVK with Edge Impulse and run local Linux inference. - [Qualcomm Dragonwing IQ-8275 Evaluation Kit](https://docs.edgeimpulse.com/hardware/boards/qualcomm-iq8275-evk.md): Set up Ubuntu on the Qualcomm Dragonwing IQ-8275 EVK, connect to Edge Impulse, and deploy accelerated models. - [Qualcomm Dragonwing IQ-9075 Evaluation Kit](https://docs.edgeimpulse.com/hardware/boards/qualcomm-iq9075-evk.md): Set up Ubuntu on the Qualcomm Dragonwing IQ-9075 EVK, connect to Edge Impulse, and deploy accelerated models. - [Qualcomm Dragonwing RB3 Gen 2 Dev Kit](https://docs.edgeimpulse.com/hardware/boards/qualcomm-rb3-gen-2-dev-kit.md): Set up the Qualcomm Dragonwing RB3 Gen 2 Dev Kit, connect to Edge Impulse, and deploy accelerated models. - [Quectel PI-SG565D](https://docs.edgeimpulse.com/hardware/boards/quectel-pi-sg565d.md): Set up the Quectel PI-SG565D for Edge Impulse, profile models, and deploy Qualcomm-accelerated inference. - [Renesas RZ/V2H](https://docs.edgeimpulse.com/hardware/boards/renesas-rz-v2h.md): Build a Renesas RZ/V2H AI SDK image, connect the board, and deploy DRP-AI accelerated models. - [Renesas RZ/V2L](https://docs.edgeimpulse.com/hardware/boards/renesas-rz-v2l.md): Build a Renesas RZ/V2L Yocto image, connect the board, and deploy DRP-AI accelerated models. - [Texas Instruments SK-AM62A-LP](https://docs.edgeimpulse.com/hardware/boards/ti-sk-am62a-lp.md): Set up the TI SK-AM62A-LP with Edge Impulse and run hardware-accelerated Linux deployments. - [Texas Instruments SK-AM68A](https://docs.edgeimpulse.com/hardware/boards/ti-sk-am68a.md): Set up the TI SK-AM68A with Edge Impulse and run hardware-accelerated Linux deployments. - [Texas Instruments SK-TDA4VM](https://docs.edgeimpulse.com/hardware/boards/ti-sk-tda4vm.md): Set up the TI SK-TDA4VM with Edge Impulse and run hardware-accelerated Linux deployments. - [Thundercomm Rubik Pi 3](https://docs.edgeimpulse.com/hardware/boards/thundercomm-rubikpi3.md): Set up Thundercomm Rubik Pi 3 on Ubuntu or Qualcomm Linux and run accelerated Edge Impulse deployments. - [Tria Vision AI-KIT 6490](https://docs.edgeimpulse.com/hardware/boards/tria-vision-ai-kit-6490.md): Set up the Tria Vision AI-KIT 6490 with Edge Impulse and deploy accelerated Linux models. - [NVIDIA Jetson](https://docs.edgeimpulse.com/hardware/boards/nvidia-jetson.md): Set up NVIDIA Jetson devices for Edge Impulse, install the Linux CLI, and deploy accelerated models. - [Deployments](https://docs.edgeimpulse.com/hardware/deployments.md): Choose a deployment path for running Edge Impulse impulses on hardware, Linux, Arduino, browsers, and other targets. - [C++ library](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-overview.md): Understand C++ signal inputs, buffer layouts, and static allocation options for running impulses in applications. - [Run C++ library on custom boards](https://docs.edgeimpulse.com/hardware/deployments/run-cpp.md): Export an Edge Impulse C++ library and integrate it into a custom board or desktop build. - [Run C++ library on desktop](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-desktop.md): Export an Edge Impulse C++ library and build a desktop application that classifies sensor data. - [Run C++ library on Android](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-android.md): Integrate an Edge Impulse Android C++ library into sample camera, static-buffer, and WearOS applications. - [Run C++ library on Alif Ensemble Series devices](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-alif-ensemble.md): Build and flash an Edge Impulse C++ library application for Alif Ensemble Series devices. - [Run C++ library on Espressif ESP-EYE (ESP32)](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-espressif-esp32.md): Build and flash an Edge Impulse C++ library application for the Espressif ESP-EYE development board. - [Run C++ library on Himax WE-I Plus](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-himax-we-i-plus.md): Build and flash an Edge Impulse C++ library application for the Himax WE-I Plus development board. - [Run C++ library on Raspberry Pi Pico (RP2040)](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-rpi-rp2040.md): Build an Edge Impulse C++ library application and load it onto a Raspberry Pi Pico board. - [Run C++ library on SiLabs Thunderboard Sense 2](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-silabs-thunderboard-sense-2.md): Import an Edge Impulse C++ library into Simplicity Studio for Thunderboard Sense 2 inference. - [Run C++ library on Sony Spresense](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-sony-spresense.md): Build and flash an Edge Impulse C++ library application for the Sony Spresense development board. - [Run C++ library on Syntiant TinyML Board](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-syntiant-tinyml-board.md): Export a Syntiant library and customize board firmware to react to keyword or motion classifications. - [Run C++ library on TI LaunchPad using GCC and the SimpleLink SDK](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-ti-launchxl.md): Build an Edge Impulse C++ library application with the SimpleLink SDK for TI LaunchPad hardware. - [Run C++ library on Zephyr-based Nordic Semiconductor development boards](https://docs.edgeimpulse.com/hardware/deployments/run-cpp-zephyr-nordic.md): Build a Zephyr RTOS application that runs an Edge Impulse C++ library on Nordic development boards. - [Run Arduino App Lab](https://docs.edgeimpulse.com/hardware/deployments/run-arduino-app-lab.md): Deploy Edge Impulse models through Arduino App Lab and run example or custom applications on UNO Q. - [Run Arduino library (IDE 2.0)](https://docs.edgeimpulse.com/hardware/deployments/run-arduino-2-0.md): Export an Arduino library for IDE 2.0, run static buffer examples, and configure supported boards. - [Run Arduino library (IDE 1.18)](https://docs.edgeimpulse.com/hardware/deployments/run-arduino-1-18.md): Export an Arduino library for IDE 1.18, run static buffer examples, and troubleshoot legacy board builds. - [Run DRP-AI library on Renesas RZ/V2L](https://docs.edgeimpulse.com/hardware/deployments/run-drpai-rzv2l.md): Download the Renesas RZ/V2L DRP-AI library and prepare a C++ application project. - [Run DRP-AI TVM i8 library on Renesas RZ/V2H](https://docs.edgeimpulse.com/hardware/deployments/run-drpai-rzv2h.md): Download the Renesas RZ/V2H DRP-AI TVM i8 library and prepare a C++ application project. - [Edge Impulse Zephyr Module Deployment](https://docs.edgeimpulse.com/hardware/deployments/run-zephyr-module.md): Run your Edge Impulse model on any Zephyr-supported board using the new Zephyr Module Deployment, no manual SDK integration required. - [Run WebAssembly library (browser)](https://docs.edgeimpulse.com/hardware/deployments/run-webassembly-browser.md): Export a WebAssembly deployment and run local impulse inference from a browser application. - [Run WebAssembly library (Node.js)](https://docs.edgeimpulse.com/hardware/deployments/run-webassembly-node.md): Export a WebAssembly deployment and run local impulse inference from a Node.js application. - [Run Arm Keil MDK CMSIS-Pack](https://docs.edgeimpulse.com/hardware/deployments/run-arm-keil-cmsis.md): Import Edge Impulse CMSIS packs into Arm Keil MDK and run standalone inference examples. - [Run Cube.MX CMSIS-Pack](https://docs.edgeimpulse.com/hardware/deployments/run-cubemx.md): Import an Edge Impulse CMSIS pack into STM32CubeMX and run inference from STM32CubeIDE projects. - [Run Docker container](https://docs.edgeimpulse.com/hardware/deployments/run-docker.md): Deploy an impulse as a Docker container with HTTP, WebSocket, offline, and hardware acceleration options. - [Run Edge Impulse firmwares](https://docs.edgeimpulse.com/hardware/deployments/run-ei-fw.md): Flash ready-to-go Edge Impulse firmware binaries to supported development boards and run the impulse runner. - [Run IAR library](https://docs.edgeimpulse.com/hardware/deployments/run-iar.md): Import an Edge Impulse library into IAR Embedded Workbench and run a standalone inference example. - [Run Linux EIM](https://docs.edgeimpulse.com/hardware/deployments/run-linux-eim.md): Download and run Edge Impulse Model artifacts for native Linux or macOS inference workflows. - [Run Open-CMSIS-Pack](https://docs.edgeimpulse.com/hardware/deployments/run-open-cmsis-pack.md): Use Open-CMSIS-Pack support to integrate Edge Impulse SDK and project packs into CMSIS-compliant tools. - [Run OpenMV library or firmware](https://docs.edgeimpulse.com/hardware/deployments/run-openmv.md): Export an impulse as an OpenMV library or firmware and run it on compatible camera boards. - [Run Particle library](https://docs.edgeimpulse.com/hardware/deployments/run-particle.md): Import a Particle library deployment into Particle Workbench and flash it to a supported board. - [Run Qualcomm IM SDK GStreamer pipeline](https://docs.edgeimpulse.com/hardware/deployments/run-qualcomm-im-sdk-gstreamer.md): Run YOLO-based Edge Impulse object detection models with Qualcomm IM SDK GStreamer pipelines. ## Tools - [Tools](https://docs.edgeimpulse.com/tools.md): Explore command-line tools, libraries, specifications, and protocols for building with Edge Impulse. - [Edge Impulse CLI](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli.md): Install and use the main CLI tools for device setup, data upload, block development, and impulse testing. - [Installation](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/installation.md): Install the Edge Impulse CLI on Linux, macOS, Windows, and WSL with troubleshooting steps. - [Serial daemon](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/serial-daemon.md): Configure supported devices over serial and proxy sensor uploads into Edge Impulse Studio. - [Uploader](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/uploader.md): Sign and upload local data files or annotated datasets to Edge Impulse projects from the CLI. - [Data forwarder](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/data-forwarder.md): Stream low-frequency sensor readings from development boards to Edge Impulse over serial. - [Impulse runner](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/impulse-runner.md): Run deployed impulses on connected devices and inspect inference output or local API responses. - [Blocks](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/blocks.md): Create, test, and publish custom organization blocks from your local development environment. - [Himax flash tool](https://docs.edgeimpulse.com/tools/clis/edge-impulse-cli/himax-flash-tool.md): Flash Edge Impulse firmware binaries to supported Himax and WiseEye devices over serial. - [Edge Impulse Linux CLI](https://docs.edgeimpulse.com/tools/clis/edge-impulse-linux-cli.md): Collect data, download EIM models, and run Linux inference tools from the command line. - [Agent Skills](https://docs.edgeimpulse.com/tools/agent-skills.md): The edgeimpulse/agent-tools repository is the public catalog of Edge Impulse Agent Skills for AI coding agents. - [C++ SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp.md): Use the portable C++ inferencing SDK to run Edge Impulse models on desktops and microcontrollers. - [Functions](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/functions.md): Reference the public C++ inferencing functions for initialization, continuous sampling, and classifier execution. - [User-defined](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/user-defined-functions.md): Implement platform hooks for timing, logging, memory allocation, sleeping, and serial I/O. - [Structs](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/structs.md): Reference C++ SDK result, bounding box, timing, and signal data structures. - [Return types](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/return-types.md): Reference Edge Impulse error codes returned by C++ inference and DSP functions. - [Macros](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/macros.md): Reference generated C++ macros that describe model dimensions, labels, inference engines, and runtime options. - [Global variables](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/cpp/global-variables.md): Reference class label arrays exposed by the generated C++ inferencing library. - [Linux SDKs](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux.md): Choose Linux SDKs and supported hardware for collecting data and running accelerated EIM models. - [Flex delegates](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/flex-delegates.md): Install and link LiteRT Flex delegates for Linux models that require unsupported native operators. - [Linux C++ SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/cpp.md): Build Linux C++ applications that collect data, classify sensor streams, and use hardware acceleration. - [Linux Go SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/go.md): Use the Linux Go SDK to collect data and classify audio, camera, or custom sensor inputs. - [Linux Node.js SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/node-js.md): Use the Linux Node.js SDK to collect data and classify audio, camera, or custom sensor inputs. - [Linux Python SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/python.md): Install the Linux Python SDK and run EIM models for audio, image, video, or custom data. - [Linux Rust SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/inference/linux/rust.md): Build Rust applications that run EIM models and upload data on Linux and macOS. - [C SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/ingestion/c.md): Use the C Ingestion SDK to sign, encode, and store sensor samples on embedded devices. - [Python SDK](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python.md): Use the Python SDK to profile, deploy, and upload Bring Your Own Model workflows programmatically. - [edgeimpulse](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/index.md) - [edgeimpulse.data](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/data/index.md) - [edgeimpulse.data.sample_type](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/data/sample_type.md) - [edgeimpulse.datasets](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/datasets/index.md) - [edgeimpulse.exceptions](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/exceptions.md) - [edgeimpulse.experimental](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/index.md) - [edgeimpulse.experimental.api](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/api.md) - [edgeimpulse.experimental.data](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/data/index.md) - [edgeimpulse.experimental.impulse](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/impulse/index.md) - [edgeimpulse.experimental.tuner](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/tuner/index.md) - [edgeimpulse.experimental.util](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/experimental/util/index.md) - [edgeimpulse.impulse](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/impulse/index.md) - [edgeimpulse.model](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/model/index.md) - [edgeimpulse.model.input_type](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/model/input_type.md) - [edgeimpulse.model.model_info](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/model/model_info.md) - [edgeimpulse.model.output_type](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/model/output_type.md) - [edgeimpulse.tuner](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/tuner/index.md) - [edgeimpulse.util](https://docs.edgeimpulse.com/tools/libraries/sdks/studio/python/edgeimpulse/util.md) - [edgeimpulse_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/index.md) - [edgeimpulse_api.api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/index.md) - [edgeimpulse_api.api.classify_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/classify_api.md) - [edgeimpulse_api.api.deployment_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/deployment_api.md) - [edgeimpulse_api.api.devices_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/devices_api.md) - [edgeimpulse_api.api.dsp_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/dsp_api.md) - [edgeimpulse_api.api.email_verification_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/email_verification_api.md) - [edgeimpulse_api.api.export_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/export_api.md) - [edgeimpulse_api.api.impulse_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/impulse_api.md) - [edgeimpulse_api.api.integrations_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/integrations_api.md) - [edgeimpulse_api.api.jobs_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/jobs_api.md) - [edgeimpulse_api.api.learn_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/learn_api.md) - [edgeimpulse_api.api.login_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/login_api.md) - [edgeimpulse_api.api.optimization_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/optimization_api.md) - [edgeimpulse_api.api.organization_blocks_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_blocks_api.md) - [edgeimpulse_api.api.organization_create_project_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_create_project_api.md) - [edgeimpulse_api.api.organization_data_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_data_api.md) - [edgeimpulse_api.api.organization_data_campaigns_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_data_campaigns_api.md) - [edgeimpulse_api.api.organization_jobs_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_jobs_api.md) - [edgeimpulse_api.api.organization_pipelines_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_pipelines_api.md) - [edgeimpulse_api.api.organization_portals_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organization_portals_api.md) - [edgeimpulse_api.api.organizations_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/organizations_api.md) - [edgeimpulse_api.api.performance_calibration_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/performance_calibration_api.md) - [edgeimpulse_api.api.post_processing_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/post_processing_api.md) - [edgeimpulse_api.api.projects_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/projects_api.md) - [edgeimpulse_api.api.raw_data_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/raw_data_api.md) - [edgeimpulse_api.api.upload_portal_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/upload_portal_api.md) - [edgeimpulse_api.api.user_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/user_api.md) - [edgeimpulse_api.api.whitelabels_api](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api/whitelabels_api.md) - [edgeimpulse_api.api_client](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/api_client.md) - [edgeimpulse_api.configuration](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/configuration.md) - [edgeimpulse_api.exceptions](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/exceptions.md) - [Python API model classes](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/models.md): Auto-generated reference stubs for the Studio Python API bindings model classes. - [edgeimpulse_api.rest](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/python/edgeimpulse_api/rest.md) - [JavaScript API Bindings](https://docs.edgeimpulse.com/tools/libraries/api-bindings/studio/javascript.md): Install, authenticate, and call Edge Impulse Studio API bindings from Node.js or browser code. - [CSV](https://docs.edgeimpulse.com/tools/specifications/data-acquisition/csv.md): Format time-series samples as CSV files for upload through Edge Impulse Studio, the CLI, or the Ingestion API. - [JSON | CBOR](https://docs.edgeimpulse.com/tools/specifications/data-acquisition/json-cbor.md): Encode signed time-series samples in the Edge Impulse data acquisition format using JSON or CBOR. - [Edge Impulse labels](https://docs.edgeimpulse.com/tools/specifications/data-annotation/ei-labels.md): Store labels and bounding boxes for exported Edge Impulse datasets using standard label files. - [Edge Impulse structured labels](https://docs.edgeimpulse.com/tools/specifications/data-annotation/ei-structured-labels.md): Define structured labels for uploading precisely segmented data files to Edge Impulse. - [Object detection](https://docs.edgeimpulse.com/tools/specifications/data-annotation/object-detection.md): Prepare image dataset annotations for Edge Impulse using bounding boxes, COCO JSON, Open Images CSV, or Pascal VOC XML. - [deployment-metadata.json](https://docs.edgeimpulse.com/tools/specifications/files/deployment-metadata-json.md): Read the metadata passed to custom deployment blocks, including impulse details, classes, DSP blocks, and deployment options. - [ei-metadata.json](https://docs.edgeimpulse.com/tools/specifications/files/ei-metadata-json.md): Update clinical dataset item metadata from a custom transformation block using the ei-metadata.json file. - [ids.json](https://docs.edgeimpulse.com/tools/specifications/files/ids-json.md): Pass sample IDs to custom AI labeling blocks so they can operate on selected data items. - [parameters.json](https://docs.edgeimpulse.com/tools/specifications/files/parameters-json.md): Describe custom block metadata, user parameters, UI controls, and conditional parameter behavior in parameters.json. - [sample_id_details.json](https://docs.edgeimpulse.com/tools/specifications/files/sample-id-details-json.md): Map learning block training and validation rows back to their original Edge Impulse sample IDs. - [train_input.json](https://docs.edgeimpulse.com/tools/specifications/files/train-input-json.md): Read the training options, dataset paths, and impulse metadata passed to custom learning blocks. - [WebSocket](https://docs.edgeimpulse.com/tools/protocols/remote-management/websocket.md): Implement the WebSocket remote management protocol for devices that exchange commands and samples with Edge Impulse. - [Serial](https://docs.edgeimpulse.com/tools/protocols/remote-management/serial.md): Implement the serial remote management protocol commands and responses for devices that connect through the Edge Impulse serial daemon. ## APIs - [APIs](https://docs.edgeimpulse.com/apis.md): Use Edge Impulse APIs to automate Studio workflows, upload data, and manage remote devices programmatically. - [Studio API](https://docs.edgeimpulse.com/apis/studio.md): Manage projects, datasets, models, jobs, and authentication through the Edge Impulse Studio API. - [Ingestion API](https://docs.edgeimpulse.com/apis/ingestion.md): Upload sensor, audio, image, and video samples to Edge Impulse through HTTP/HTTPS ingestion endpoints. - [Remote Management API](https://docs.edgeimpulse.com/apis/remote-management.md): Connect devices to Edge Impulse Studio through the remote management service and WebSocket protocol. ## Tutorials - [Tutorials](https://docs.edgeimpulse.com/tutorials.md): Explore step-by-step tutorials for building, deploying, and integrating Edge Impulse models with devices and data. - [End-to-end tutorials](https://docs.edgeimpulse.com/tutorials/end-to-end/overview.md): Choose an end-to-end tutorial for image, audio, sensor, industrial, or platform-specific machine learning projects. - [Image classification](https://docs.edgeimpulse.com/tutorials/end-to-end/image-classification.md): Train a transfer-learning model to recognize objects and deploy it to a microcontroller. - [Object detection with bounding boxes](https://docs.edgeimpulse.com/tutorials/end-to-end/object-detection-bounding-boxes.md): Detect and localize multiple objects in a scene with bounding boxes using transfer learning on an edge device. - [Object detection with centroids](https://docs.edgeimpulse.com/tutorials/end-to-end/object-detection-centroids.md): Detect, count, and track objects in real time on a microcontroller using FOMO, 30× faster than MobileNet SSD and under 200K RAM. - [Visual anomaly detection](https://docs.edgeimpulse.com/tutorials/end-to-end/visual-anomaly-detection.md): Learn what 'normal' looks like and automatically detect defects or unusual patterns in images without labeled anomaly data. - [Visual regression](https://docs.edgeimpulse.com/tutorials/end-to-end/visual-regression.md): Build a model that predicts a continuous value, like a fill level or angle, directly from images and deploy it to an embedded device. - [Keyword spotting](https://docs.edgeimpulse.com/tutorials/end-to-end/keyword-spotting.md): Build a voice keyword detection system that runs entirely on an embedded device. - [Sound recognition](https://docs.edgeimpulse.com/tutorials/end-to-end/sound-recognition.md): Build a model that recognizes specific sounds, like running water, from audio and deploy it to an embedded device. - [Motion recognition with anomaly detection](https://docs.edgeimpulse.com/tutorials/end-to-end/motion-recognition.md): Recognize gestures from accelerometer data with a neural network classifier and anomaly detection, deployed to a microcontroller. - [Environmental sensor fusion](https://docs.edgeimpulse.com/tutorials/end-to-end/environmental-sensor-fusion.md): Fuse accelerometer and temperature sensor data to build a multi-sensor classifier, and deploy it to an Arduino Nano 33 BLE Sense. - [HR/HRV estimation](https://docs.edgeimpulse.com/tutorials/end-to-end/hr-hrv-estimation.md): Extract heart rate and heart rate variability (HRV) features from PPG data to power health-monitoring applications at the edge. - [Temperature anomaly detection on Opta PLC](https://docs.edgeimpulse.com/tutorials/end-to-end/temp-anomaly-detection-opta-plc.md): Detect temperature anomalies on an Arduino Opta PLC using Edge Impulse, with optional Arduino Cloud integration for remote monitoring. - [OBD Automotive data](https://docs.edgeimpulse.com/tutorials/end-to-end/obd-automotive-data.md): Process OBD-II and CAN bus time-series data to classify vehicle faults, and run inference on a Linux SBC. - [Avnet RASynBoard - motion recognition](https://docs.edgeimpulse.com/tutorials/hardware/avnet-rasyn-motion-recognition.md): Recognize gestures with a neural network on the Avnet RASynBoard (Renesas RA6 + Syntiant NDP 120) using Edge Impulse. - [Particle Photon 2 - audio event detection](https://docs.edgeimpulse.com/tutorials/hardware/particle-photon2-audio-even-detection.md): Detect audio events like a doorbell ring on the Particle Photon 2 using a microphone and Edge Impulse. - [Particle Photon 2 and Boron - motion recognition](https://docs.edgeimpulse.com/tutorials/hardware/particle-photon2-motion-recognition.md): Recognize gestures from IMU data on the Particle Photon 2 or Boron using Edge Impulse. - [Renesas RA8D1 - NVIDIA TAO models](https://docs.edgeimpulse.com/tutorials/hardware/renesas-ra8d1-nvidia-tao.md): Run NVIDIA TAO pre-trained object detection and classification models on the Renesas RA8D1 Cortex-M85 MCU. - [Renesas RZ/V2L - object detection (x2)](https://docs.edgeimpulse.com/tutorials/hardware/renesas-rzv2l-object-detection-x2.md): Run multiple object detection models simultaneously across two cameras on the Renesas RZ/V2L with DRPAI acceleration. - [Sony Spresense - image classification](https://docs.edgeimpulse.com/tutorials/hardware/sony-spresense-image-classification.md): Classify objects through a camera using transfer learning on the Sony Spresense, and deploy the model to the device. - [SiLabs xG24 Dev Kit - object detection](https://docs.edgeimpulse.com/tutorials/hardware/silabs-xg24-devkit-object-detection.md): Count people in real time using FOMO object detection on the SiLabs xG24 Dev Kit with AI/ML hardware accelerator. - [Syntiant NDPs - keyword spotting](https://docs.edgeimpulse.com/tutorials/hardware/syntiant-ndp-keyword-spotting.md): Build a keyword spotting model and deploy it to Syntiant TinyML, Arduino Nicla Voice, or Avnet RASynBoard using Edge Impulse. - [Syntiant NDPs - motion recognition](https://docs.edgeimpulse.com/tutorials/hardware/syntiant-ndp-motion-recognition.md): Recognize circular gestures from IMU data on the Syntiant TinyML Board or Arduino Nicla Voice using Edge Impulse. - [TI LaunchXL - keyword spotting](https://docs.edgeimpulse.com/tutorials/hardware/ti-launchxl-keyword-spotting.md): Detect voice keywords on the TI LaunchXL using a deep neural network classifier trained in Edge Impulse. - [TI LaunchXL - sound recognition](https://docs.edgeimpulse.com/tutorials/hardware/ti-launchxl-sound-recognition.md): Recognize specific sounds like running water on the TI LaunchXL using an audio classification model from Edge Impulse. - [Motion recognition using an agent](https://docs.edgeimpulse.com/tutorials/end-to-end/motion-recognition-using-agent.md): Use an AI coding agent to take an Edge Impulse project from raw data collection through model training and deployment to a device. - [Install Agent Skills](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/create-edge-impulse-skill.md): Install Agent Skills from the Edge Impulse agent-tools repository to give AI coding agents persistent knowledge of the Edge Impulse platform, APIs, and deployment workflows. - [Create a companion skill](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/create-a-companion-skill.md): Use a companion skill that knows your target hardware and SDK so an AI agent can write application code that runs the exported model on your device. - [Prompt library](https://docs.edgeimpulse.com/tutorials/topics/ai-agents/prompt-library.md): Ready-to-use AI agent prompts for every stage of an Edge Impulse project. - [Collect image data using a mobile phone](https://docs.edgeimpulse.com/tutorials/topics/data/collect-image-data-phone.md): Collect and import labeled image data into Edge Impulse using your mobile phone as a wireless camera. - [Collect image data using Studio](https://docs.edgeimpulse.com/tutorials/topics/data/collect-image-data-studio.md): Collect labeled image data directly into Edge Impulse Studio using a connected development board with an integrated camera. - [Collect image data using OpenMV Cam H7 Plus](https://docs.edgeimpulse.com/tutorials/topics/data/collect-image-data-openmv-h7-plus.md): Collect and import labeled image data into Edge Impulse using the OpenMV Cam H7 Plus camera module. - [Generate image data using Dall-E](https://docs.edgeimpulse.com/tutorials/topics/data/generate-image-data-dall-e.md): Use OpenAI's DALL-E to generate synthetic image datasets for Edge Impulse when real-world data is unavailable or difficult to collect. - [Generate image data using RealVisXL](https://docs.edgeimpulse.com/tutorials/topics/data/generate-image-data-realvisxl.md): Generate photorealistic synthetic image datasets for Edge Impulse using the RealVisXL model hosted on Hugging Face. - [Generate keyword data using Google TTS](https://docs.edgeimpulse.com/tutorials/topics/data/generate-keyword-data-google-tts.md): Generate synthetic keyword spotting audio datasets for Edge Impulse using Google Text-to-Speech to programmatically create labeled voice samples. - [Generate time-series data using MATLAB](https://docs.edgeimpulse.com/tutorials/topics/data/generate-time-series-data-matlab.md): Generate synthetic time-series motion data in MATLAB using the Signal Processing Toolbox, and upload it to Edge Impulse for training. - [Generate time-series data using PyBullet](https://docs.edgeimpulse.com/tutorials/topics/data/generate-time-series-data-pybullet.md): Generate synthetic accelerometer time-series datasets using PyBullet physics simulation and upload them to Edge Impulse for model training. - [Label audio data using your existing model](https://docs.edgeimpulse.com/tutorials/topics/data/label-audio-data-existing-model.md): Use an existing Edge Impulse model to run inference on raw audio data and automatically upload the labeled results back to your project. - [Label image data using GPT-4o](https://docs.edgeimpulse.com/tutorials/topics/data/label-image-data-gpt-4o.md): Automatically label unlabeled image datasets in Edge Impulse using GPT-4o vision to classify images based on a custom text prompt. - [Build a custom processing block](https://docs.edgeimpulse.com/tutorials/topics/feature-extraction/build-custom-processing-blocks.md): Extend Edge Impulse with your own signal processing logic by building and deploying a custom processing block to Studio. - [Use embeddings for sensor fusion](https://docs.edgeimpulse.com/tutorials/topics/feature-extraction/use-embeddings-sensor-fusion.md): Combine high-dimensional audio or image embeddings with time-series sensor data for advanced sensor fusion using custom DSP blocks. - [Classify multiple 2D input features](https://docs.edgeimpulse.com/tutorials/topics/machine-learning/classify-multiple-2d-features.md): Train a neural network with multiple 2D feature branches for sensor fusion classification using a single Edge Impulse model. - [Deploy pretrained models with freeform outputs](https://docs.edgeimpulse.com/tutorials/topics/machine-learning/deploy-freeform-model.md): Import and deploy a pretrained model with custom input/output formats into Edge Impulse using freeform model deployment. - [Use the regression learning block](https://docs.edgeimpulse.com/tutorials/topics/machine-learning/use-regression-learning-block.md): Design a regression model in Edge Impulse to predict continuous values from time-series sensor data, applied to a cold-chain monitoring use case. - [Visualize neural networks decisions with Grad-CAM](https://docs.edgeimpulse.com/tutorials/topics/machine-learning/visualize-neural-netowrks-gradcam.md): Use Grad-CAM to visualize which parts of an image your neural network focuses on when making classification decisions. - [Run multiple impulses (C++)](https://docs.edgeimpulse.com/tutorials/topics/inference/run-multiple-impulses-cpp.md): Run two independent Edge Impulse impulses simultaneously on a microcontroller using the C++ SDK. - [Run on Qualcomm Device Cloud](https://docs.edgeimpulse.com/tutorials/topics/inference/run-qualcomm-device-cloud.md): Run Edge Impulse inference on real Qualcomm hardware remotely using Qualcomm Device Cloud, without needing to purchase physical devices. - [Sample audio continuously](https://docs.edgeimpulse.com/tutorials/topics/inference/sample-audio-continuously.md): Continuously capture and classify audio on an embedded device using the Edge Impulse SDK's continuous inferencing mode to avoid missing audio events. - [Count objects using FOMO](https://docs.edgeimpulse.com/tutorials/topics/post-processing/count-objects-fomo.md): Post-process FOMO detections to accurately count objects by tracking centroids across video frames. - [Two-stage OCR on Linux devices](https://docs.edgeimpulse.com/tutorials/topics/model-cascades/ocr.md): Build a two-stage OCR pipeline on Linux edge devices using model cascading: detect text regions, then recognize characters. - [Acquire and process data using Golioth on AI](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/acquire-process-golioth.md): Securely collect and process labeled sensor data from cellular-connected devices using Golioth on AI and Edge Impulse custom transformation blocks. - [Deploy using GitHub Actions](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/deploy-github-actions.md): Automate Edge Impulse model builds and deployments using GitHub Actions and the Edge Impulse REST API for continuous integration and delivery. - [Allxon](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-allxon.md): Deploy over-the-air Edge Impulse model updates to NVIDIA Jetson edge devices using Allxon's remote device management platform. - [Arduino IDE (for ESP32)](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-arduino.md): Deploy over-the-air (OTA) Edge Impulse model updates to ESP32 devices using the Arduino IDE and the Edge Impulse REST API for automated update checks. - [Arduino IoT Cloud](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-arduino-iot-cloud.md): Deploy over-the-air (OTA) Edge Impulse impulse updates to Arduino devices using the Arduino IoT Cloud. - [Balena](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-balena.md): Deploy over-the-air (OTA) Edge Impulse model updates to Linux edge devices using Balena's container-based fleet management platform. - [Blues Wireless](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-blues-wireless.md): Deploy over-the-air (OTA) Edge Impulse firmware updates to Blues Wireless Notecard-connected devices using Notehub DFU. - [C++](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-cpp.md): Implement over-the-air (OTA) Edge Impulse model updates in C++ by storing the model in a file or partition separate from the application binary. - [Espressif IDF](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-espressif-idf.md): Deploy over-the-air (OTA) Edge Impulse model updates to ESP32 devices using the Espressif IDF OTA framework and Edge Impulse REST API. - [Nordic Thingy53](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-thingy53.md): Deploy over-the-air (OTA) Edge Impulse model updates to the Nordic Thingy:53 using the Nordic Thingy app and the firmware DFU pipeline. - [NVIDIA Jetson](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-jetson-nano.md): Deploy over-the-air (OTA) Edge Impulse model updates to NVIDIA Jetson Nano using Docker containers and the Edge Impulse Linux runner. - [Particle Workbench](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-particle-workbench.md): Deploy over-the-air (OTA) Edge Impulse impulse updates to Particle Photon 2 devices using Particle Workbench and the Edge Impulse API. - [Particle Webhook](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-particle-webhook.md): Configure a Particle webhook to forward sensor data to the Edge Impulse ingestion API as part of an OTA model update and active learning pipeline. - [Zephyr on Golioth](https://docs.edgeimpulse.com/tutorials/topics/lifecycle-management/ota-zephyr-golioth.md): Deploy over-the-air (OTA) Edge Impulse model updates to Zephyr RTOS devices using Golioth's cloud OTA service and the Edge Impulse REST API. - [Android series overview](https://docs.edgeimpulse.com/tutorials/topics/android/android-series.md): Deploy Edge Impulse models on Android using the NDK and Edge Impulse C++ SDK - [Static buffer inference on Android](https://docs.edgeimpulse.com/tutorials/topics/android/static-buffer-inference.md): Run Edge Impulse inference with pre-loaded test data on Android - [Keyword spotting on Android](https://docs.edgeimpulse.com/tutorials/topics/android/keyword-spotting.md): Build a real-time audio keyword recognition app on Android using your phone's microphone - [QNN hardware acceleration on Android](https://docs.edgeimpulse.com/tutorials/topics/android/qnn-acceleration.md): Enable Qualcomm AI Engine Direct SDK acceleration for Edge Impulse models on Snapdragon devices - [Android data collector](https://docs.edgeimpulse.com/tutorials/topics/android/android-data-collector.md): Collect phone sensor data, images, and BLE-relayed inference results from a Zephyr device, and upload directly to your Edge Impulse project - [AR/XR factory floor (Unity + Meta Quest 3)](https://docs.edgeimpulse.com/tutorials/topics/android/ar-xr-factory-floor.md): Run an Edge Impulse FOMO vision model fully on-device inside a Unity XR app on Meta Quest 3 / 3S - [Zephyr series overview](https://docs.edgeimpulse.com/tutorials/topics/zephyr/zephyr-module-series.md): A series of tutorials on using Edge Impulse Zephyr Module for deploying machine learning models using the Edge Impulse Zephyr Module - [Porting Zephyr Projects Between Boards](https://docs.edgeimpulse.com/tutorials/topics/zephyr/porting-between-boards.md): Learn how to port Edge Impulse Zephyr applications between different hardware platforms - [IMU inference on Zephyr](https://docs.edgeimpulse.com/tutorials/topics/zephyr/zephyr-module-imu.md): Run Edge Impulse machine learning models for real-time motion classification on Zephyr RTOS using IMU sensors - [Microphone Keyword Spotting on Zephyr](https://docs.edgeimpulse.com/tutorials/topics/zephyr/zephyr-module-microphone-kws.md): Run Edge Impulse audio keyword spotting models on Zephyr RTOS using PDM microphones - [BLE GATT sensor bridge for Edge Impulse](https://docs.edgeimpulse.com/tutorials/topics/zephyr/zephyr-ble-gatt-client.md): Stream IMU data and inference results over BLE with Zephyr for Nesso N1, Arduino Nano 33 BLE Sense, and Nordic Thingy:53. - [Robotics with Edge Impulse and ROS 2](https://docs.edgeimpulse.com/tutorials/topics/robotics-series/robotics-series.md): Build robotics workflows that combine ROS 2, Edge Impulse vision models, simulation, and physical robot deployment. - [Run Edge Impulse Object Detection with ROS 2](https://docs.edgeimpulse.com/tutorials/topics/robotics-series/edgeimpulse-ros2-integration.md): Use edgeimpulse_ros to run a Linux .eim model on a ROS 2 image topic and publish vision_msgs detection messages. - [Run Edge Impulse Vision on a JetBot with Rubik Pi](https://docs.edgeimpulse.com/tutorials/topics/robotics-series/wheeled-robot-ros2-integration.md): Combine JetBot ROS 2 motor control on Qualcomm Rubik Pi with Edge Impulse Linux object detection. - [Run Edge Impulse Pick and Place with UNO Q Braccio](https://docs.edgeimpulse.com/tutorials/topics/robotics-series/robotic-arm-ros2-integration.md): Use Edge Impulse object detection, ROS 2, and an Arduino UNO Q to sort colored blocks with a TinkerKit Braccio arm. - [Build a Mobile Inspection System with Edge Impulse and ROS 2](https://docs.edgeimpulse.com/tutorials/topics/robotics-series/wheeled-inspection-system.md): Design a wheeled rover inspection workflow with ROS 2, Edge Impulse detection, anomaly scoring, gimbal control, logging, and safety layers. - [Ingest structured label data](https://docs.edgeimpulse.com/tutorials/tools/apis/ingestion/ingest-structured-label-data.md): Upload and update time-series samples with structured labels through the Edge Impulse Ingestion API. - [Collect data from a device](https://docs.edgeimpulse.com/tutorials/tools/apis/studio/collect-data-device.md): Use the Studio API to trigger data sampling from a connected device and collect new project data. - [Customize the EON Tuner](https://docs.edgeimpulse.com/tutorials/tools/apis/studio/customize-the-eon-tuner.md): Configure, start, and monitor EON Tuner jobs programmatically with the Edge Impulse Studio API. - [Deploy a model](https://docs.edgeimpulse.com/tutorials/tools/apis/studio/deploy-model.md): Start a deployment build with the Studio API, monitor the job, and download the generated model archive. - [Use the Python API bindings](https://docs.edgeimpulse.com/tutorials/tools/api-bindings/studio/python/use-python-api-bindings.md): Use the Python API bindings to upload data, create an impulse, train, test, and deploy a project. - [Use the Edge Impulse Python SDK to upload and download data](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/upload-download-data.md): Upload, download, and delete image, object detection, CSV, JSON, NumPy, and pandas samples with the Python SDK. - [Use the Edge Impulse Python SDK to run the EON Tuner](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/run-eon-tuner.md): Upload data, run the EON Tuner, review trials, and deploy the best impulse with the Python SDK. - [Use the Edge Impulse Python SDK with Hugging Face](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/use-hugging-face.md): Download an ONNX model from Hugging Face, profile it, and deploy it as an Edge Impulse C++ library. - [Use the Edge Impulse Python SDK with SageMaker Studio](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/use-sagemaker-studio.md): Train an image classifier in SageMaker Studio, convert it to TensorFlow Lite, and deploy it with Edge Impulse. - [Use the Edge Impulse Python SDK with TensorFlow and Keras](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/use-tf-keras.md): Train a TensorFlow and Keras MNIST model, profile it with the Python SDK, and export a C++ deployment. - [Use the Edge Impulse Python SDK with Weights & Biases](https://docs.edgeimpulse.com/tutorials/tools/sdks/studio/python/use-wandb.md): Track TensorFlow experiments in Weights & Biases while profiling edge performance with the Edge Impulse Python SDK. - [Arduino Machine Learning Tools](https://docs.edgeimpulse.com/tutorials/integrations/arduino-ml-tools.md): Use Arduino Machine Learning Tools to create projects, deploy models, and share Edge Impulse workflows from Arduino Cloud. - [AWS IoT Greengrass](https://docs.edgeimpulse.com/tutorials/integrations/aws-greengrass.md): Deploy Edge Impulse Linux runner components with AWS IoT Greengrass, S3 artifacts, Secrets Manager, and IoT Core topics. - [Deploy PyTorch models with ExecuTorch](https://docs.edgeimpulse.com/tutorials/integrations/executorch.md): Train PyTorch models in Edge Impulse and export ExecuTorch programs using custom learning and deployment blocks - [AngleForge: consistent image angles with Qwen Image Edit](https://docs.edgeimpulse.com/tutorials/integrations/huggingface-angleforge-synthetic-data.md): Normalize project images to consistent camera angles with AngleForge synthetic data blocks and the Hugging Face Space. - [NVIDIA Omniverse](https://docs.edgeimpulse.com/tutorials/integrations/nvidia-omniverse.md): Upload synthetic computer vision datasets from NVIDIA Omniverse to Edge Impulse and run local classification in Omniverse. - [Scailable](https://docs.edgeimpulse.com/tutorials/integrations/scailable.md): Import Edge Impulse vision models into Scailable Cloud Platform and deploy them to supported edge AI accelerators. - [syntheticAIdata Vision Datasets](https://docs.edgeimpulse.com/tutorials/integrations/syntheticaIdata-vision-datasets.md): Upload synthetic Vision Datasets images to Edge Impulse for computer vision training, verification, and troubleshooting. - [Weights & Biases](https://docs.edgeimpulse.com/tutorials/integrations/weights-and-biases.md): Track Edge Impulse model metrics in Weights & Biases and follow W&B resources for custom blocks and sweeps. - [ROS 2 (edgeimpulse_ros)](https://docs.edgeimpulse.com/tutorials/integrations/ros2.md): Run Edge Impulse Linux image models in ROS 2 nodes that publish vision messages for robotic perception. ## Projects - [Projects](https://docs.edgeimpulse.com/projects.md): Explore Expert Network projects that combine Edge Impulse models with hardware, software, and real-world deployments. - [Project list](https://docs.edgeimpulse.com/projects/expert-network/project-list.md): Browse Expert Network projects by use case, hardware target, and machine learning workflow. - [Mini-figurine Cataloger and Listing Tracker with Hermes Agent](https://docs.edgeimpulse.com/projects/expert-network/mini-figurine-cataloger-and-listing-tracker-with-hermes-agent.md): Build an Arduino UNO Q cataloger rig that photographs mini-figurines and tracks listings with Edge Impulse and Hermes. - [Hand Gestures Control for Robotic Arm on Arduino UNO Q](https://docs.edgeimpulse.com/projects/expert-network/hand-gestures-control-robotic-arm-arduino-uno-q.md): Control an Arduino UNO Q robotic arm with hand gesture detection and Arduino App Lab. - [Instance Segmentation on Edge Impulse with a Two-Stage Model Cascade](https://docs.edgeimpulse.com/projects/expert-network/instance-seg-byom-freeform-person-blur.md): Cascade Edge Impulse object detection with BYOM Freeform instance segmentation to blur people on Rubik Pi 3. - [Model Cascading from Visual Anomaly Detection to VLM with Arduino App Lab](https://docs.edgeimpulse.com/projects/expert-network/vision-language-models-advancing-production-inspection.md): Cascade visual anomaly detection and a SmolVLM model on Arduino UNO Q for package inspection. - [Multimodal AI Asset Protection Prototype](https://docs.edgeimpulse.com/projects/expert-network/multimodal-ai-asset-protection-prototype.md): Protect a valuable object with a Rubik Pi 3 prototype that combines vision, environmental sensors, and Telegram alerts. - [Multilayer AI Agent Drone Detection](https://docs.edgeimpulse.com/projects/expert-network/multilayer-ai-agent-drone-detection-rubik-pi-3.md): Build a Rubik Pi 3 drone detection agent that combines audio classification, camera input, and Telegram alerts. - [From TinyML to VLMs: Bringing Visual Intelligence to App Lab](https://docs.edgeimpulse.com/projects/expert-network/from-tinyml-to-vlms-bringing-visual-intelligence-to-app-lab.md): Cascade a FOMO detector with local SmolVLM inference in Arduino App Lab on UNO Q. - [Face-following robot with Arduino UNO Q and Edge Impulse](https://docs.edgeimpulse.com/projects/expert-network/face-following-robot-uno-q.md): Build an Arduino UNO Q robot that uses face detection, Bridge RPC, and a web UI for live steering. - [AI-driven Web-based Ancillary Lab Assistant | UNO Q and Gemini](https://docs.edgeimpulse.com/projects/expert-network/ai-driven-ancillary-lab-assistant-uno-q-gemini.md): Build an Arduino UNO Q lab assistant with Edge Impulse object detection, sensor dashboards, and Google Gemini lessons. - [Object Following Pan-Tilt Camera System - Rubik Pi 3](https://docs.edgeimpulse.com/projects/expert-network/face-tracking-yolo-pro-rubik-pi.md): Build a Rubik Pi 3 pan-tilt camera that tracks faces and gestures with a YOLO-Pro model. - [Audio Classification and GPIO Response - Rubik Pi 3](https://docs.edgeimpulse.com/projects/expert-network/audio-classification-gpio-rubik-pi.md): Train glass-breaking audio classification on Rubik Pi 3 and trigger GPIO responses from inference results. - [YOLO-Pro Smart Parking Meter - Rubik Pi 3](https://docs.edgeimpulse.com/projects/expert-network/smart-parking-meter-rubik-pi.md): Train and deploy a YOLO-Pro parking meter model on Rubik Pi 3 with hardware-accelerated inference. - [Anomaly Detection with the Rubik Pi 3 and n8n](https://docs.edgeimpulse.com/projects/expert-network/anomaly-detection-n8n-rubik-pi.md): Detect visual anomalies on a Rubik Pi 3 with Edge Impulse and report results through an n8n workflow. - [Vehicle Security Camera - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/vehicle-security-camera-arduino-portenta-h7.md): Build an Arduino Portenta H7 surveillance camera that classifies suspicious activity around parked vehicles. - [Water Pollution Detection - Arduino Nano ESP32 + Ultrasonic Scan](https://docs.edgeimpulse.com/projects/expert-network/water-pollution-detection-arduino-nano-esp32.md): Build a dual-board AIoT system that uses ultrasonic scans and water tests to detect pollution risks. - [Renesas CK-RA6M5 Cloud Kit - Getting Started with Machine Learning](https://docs.edgeimpulse.com/projects/expert-network/renesas-ra6m5-getting-started.md): Connect the Renesas CK-RA6M5 Cloud Kit to Edge Impulse and deploy a local audio classifier. - [TI CC1352P Launchpad - Getting Started with Machine Learning](https://docs.edgeimpulse.com/projects/expert-network/ti-cc1352p-getting-started.md): Collect accelerometer data on the TI CC1352P LaunchPad and deploy a motion classifier. - [OpenMV Cam RT1062 - Getting Started with Machine Learning](https://docs.edgeimpulse.com/projects/expert-network/getting-started-openmv-rt1062.md): Evaluate OpenMV Cam RT1062 image classification performance, power use, and deployment with Edge Impulse. - [Getting Started with Edge Impulse Experiments](https://docs.edgeimpulse.com/projects/expert-network/experiments-getting-started.md): Use Edge Impulse Experiments and EON Tuner to compare multiple impulses on Arduino vibration data. - [Workplace Organizer - Nvidia Jetson Nano](https://docs.edgeimpulse.com/projects/expert-network/workplace-organizer-nvidia-jetson-nano.md): Detect missing workplace tools with colored markers, a Jetson Nano, and an Edge Impulse object detection model. - [Recyclable Materials Sorter - Nvidia Jetson Nano](https://docs.edgeimpulse.com/projects/expert-network/recyclable-materials-sorter-nvidia-jetson-nano.md): Train and deploy a Jetson Nano image classifier to sort glass, plastic, and aluminum containers. - [Analog Meter Reading - Arduino Nicla Vision](https://docs.edgeimpulse.com/projects/expert-network/analog-meter-reading-arduino-nicla-vision.md): Train and deploy an Arduino Nicla Vision image classifier that reads analog boiler pressure gauges from generated datasets. - [Creating Synthetic Data with Nvidia Omniverse Replicator](https://docs.edgeimpulse.com/projects/expert-network/nvidia-omniverse-replicator.md): Generate randomized synthetic cutlery data with NVIDIA Omniverse Replicator and train an Edge Impulse object detector. - [SonicSight AR - Sound Classification with Feedback on an Augmented Reality Display](https://docs.edgeimpulse.com/projects/expert-network/particle-photon-2-sonicsight-ar.md): Classify safety-related sounds with Particle Photon 2 and show alerts on a Brilliant Labs Monocle display. - [Traffic Monitoring - Brainchip Akida](https://docs.edgeimpulse.com/projects/expert-network/brainchip-akida-traffic-monitoring.md): Build a BrainChip Akida traffic monitoring application with object detection, deployment, and live inference. - [Multi-camera Video Stream Inference - Brainchip Akida](https://docs.edgeimpulse.com/projects/expert-network/brainchip-akida-multi-camera-inference.md): Run multi-camera video stream inference on BrainChip Akida using Edge Impulse object detection models. - [Industrial Inspection Line - Brainchip Akida](https://docs.edgeimpulse.com/projects/expert-network/brainchip-akida-industrial-inspection.md): Deploy an Akida object detection workflow for industrial inspection and run it in a desktop application. - [X-Ray Classification and Analysis - Brainchip Akida](https://docs.edgeimpulse.com/projects/expert-network/brainchip-akida-xray-classification.md): Classify X-ray images on a BrainChip Akida developer kit and compare Python and webcam inference methods. - [Inventory Stock Tracker - FOMO - BrainChip Akida](https://docs.edgeimpulse.com/projects/expert-network/fomo-stock-tracker-brainchip.md): Track inventory parts with FOMO on Raspberry Pi 5, BrainChip Akida, and an LED matrix. - [Container Counting - Arduino Nicla Vision](https://docs.edgeimpulse.com/projects/expert-network/container-counting-arduino-nicla-vision.md): Train and deploy a FOMO object detection model on Arduino Nicla Vision to count warehouse containers. - [Smart Smoke Alarm - Arduino Nano 33](https://docs.edgeimpulse.com/projects/expert-network/smart-smoke-alarm-arduino-nano-33.md): Detect people in a simulated smoke alarm with thermal images, Edge Impulse, and Arduino hardware. - [Shield Bot Autonomous Security Robot](https://docs.edgeimpulse.com/projects/expert-network/shieldbot-autonomous-security-robot.md): Build a Raspberry Pi security robot that detects anomalous sounds, patrols, and identifies people with Edge Impulse models. - [Cyclist Blind Spot Detection - Himax WE-I Plus](https://docs.edgeimpulse.com/projects/expert-network/cyclist-blind-spot-detection-himax-we-i-plus.md): Train and deploy a FOMO cyclist detector on the Himax WE-I Plus to prototype blind spot warnings. - [IV Drip Fluid-Level Monitoring - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/iv-drip-fluid-level-monitoring-arduino-portenta-h7.md): Monitor IV drip fluid levels with computer vision on Arduino Portenta H7 and deploy the model to a device. - [Worker PPE Safety Monitoring - Nvidia Jetson Nano](https://docs.edgeimpulse.com/projects/expert-network/worker-safety-monitoring-nvidia-jetson-nano.md): Train a FOMO object detection model on a NVIDIA Jetson Nano to identify missing PPE in workplace images. - [Delivered Package Detection - ESP-EYE](https://docs.edgeimpulse.com/projects/expert-network/delivered-package-detection-esp-eye.md): Detect delivered parcels with FOMO on ESP-EYE and trigger follow-up actions for package monitoring. - [Bean Leaf Disease Classification - Sony Spresense](https://docs.edgeimpulse.com/projects/expert-network/bean-leaf-disease-classification-sony-spresense.md): Classify bean leaf diseases with a Sony Spresense camera, Edge Impulse training, and on-device deployment. - [Oil Tank Measurement Using Computer Vision - Sony Spresense](https://docs.edgeimpulse.com/projects/expert-network/oil-tank-gauge-monitoring-sony-spresense.md): Read analog heating oil tank gauges with visual regression on Sony Spresense and send readings to AWS IoT. - [Object Counting for Smart Industries - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/object-counting-for-smart-industries-raspberry-pi.md): Build a Raspberry Pi vision counter with FOMO, Firebase, and a web dashboard for industrial parts. - [Smart Cashier with FOMO - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/smart-cashier-with-fomo-raspberry-pi.md): Use FOMO on a Raspberry Pi to identify cashier-table products and calculate totals without barcodes. - [PCB Defect Detection with Computer Vision - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/pcb-defect-detection-with-computer-vision-raspberry-pi.md): Train FOMO to detect missing holes, open circuits, and short circuits on PCBs before Raspberry Pi deployment. - [Bicycle Counting - Sony Spresense](https://docs.edgeimpulse.com/projects/expert-network/bicycle-counting-sony-spresense.md): Count bicycles from a Sony Spresense camera using computer vision and a lightweight roadside sensor. - [Counting Eggs with Computer Vision - OpenMV Cam H7](https://docs.edgeimpulse.com/projects/expert-network/counting-eggs-openmv-cam-h7.md): Build an OpenMV Cam H7 system that counts quail eggs, tracks feeder status, and sends WhatsApp alerts. - [Elevator Passenger Counting - Arduino Nicla Vision](https://docs.edgeimpulse.com/projects/expert-network/elevator-passenger-counting-arduino-nicla-vision.md): Count elevator passengers with Arduino Nicla Vision and log usage data for overload alerts and reporting. - [ESD Protection using Computer Vision - Seeed ReComputer](https://docs.edgeimpulse.com/projects/expert-network/esd-protection-using-computer-vision-seeed-recomputer.md): Detect ESD handling risks on a Seeed reComputer with FOMO object detection for wrist straps, rings, and watches. - [Solar Panel Defect Detection - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/solar-panel-defect-detection-arduino-portenta-h7.md): Build a FOMO model on Arduino Portenta H7 to detect cracks in solar panel images. - [Label Defect Detection - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/label-defect-detection-raspberry-pi.md): Detect bottle label defects with a Raspberry Pi camera, Edge Impulse vision model, and Firebase dashboard. - [Dials and Knob Monitoring with Computer Vision - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/dials-and-knob-monitoring-with-computer-vision-raspberry-pi.md): Monitor analog console knobs with Raspberry Pi computer vision and stop recordings when controls move out of range. - [Digital Character Recognition on Electric Meter System - OpenMV Cam H7](https://docs.edgeimpulse.com/projects/expert-network/digital-character-recognition-on-electric-meter-system-openmv-cam-h7.md): Recognize electric meter digits with an OpenMV camera and publish readings through a low-power IoT workflow. - [Corrosion Detection with Computer Vision - Seeed reTerminal](https://docs.edgeimpulse.com/projects/expert-network/corrosion-detection-computer-vision-seeed-reterminal.md): Train and deploy a computer vision model on Seeed reTerminal to classify rust and no-rust images. - [Inventory Management with Computer Vision - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/inventory-management-with-computer-vision-raspberry-pi.md): Build a Raspberry Pi computer vision system that counts inventory items and updates Firebase for a web dashboard. - [Monitoring Retail Checkout Lines with Computer Vision - Renesas RZ/V2L](https://docs.edgeimpulse.com/projects/expert-network/renesas-rzv2l-monitoring-retail-checkout-lines.md): Build a Renesas RZ/V2L computer vision application that detects queue lengths at retail checkout counters. - [Counting Retail Inventory with Computer Vision - Renesas RZ/V2L](https://docs.edgeimpulse.com/projects/expert-network/renesas-rzv2l-counting-retail-inventory.md): Train and deploy a Renesas RZ/V2L object detection model that counts retail shelf inventory. - [Pose Detection - Renesas RZ/V2L](https://docs.edgeimpulse.com/projects/expert-network/renesas-rzv2l-pose-detection.md): Learn how Renesas DRP-AI acceleration works with Edge Impulse deployment options for vision and pose models. - [Product Quality Inspection - Renesas RZ/V2L](https://docs.edgeimpulse.com/projects/expert-network/renesas-rzv2l-product-quality-inspection.md): Deploy a Renesas RZ/V2L object detection application that inspects submersible pump impellers on a conveyor. - [Smart Grocery Cart Using Computer Vision - OpenMV Cam H7](https://docs.edgeimpulse.com/projects/expert-network/smart-grocery-cart-with-computer-vision-openmv-cam-h7.md): Build a grocery cart prototype that detects products with OpenMV, tracks purchases, and emails checkout details. - [Driver Drowsiness Detection With FOMO - Arduino Nicla Vision](https://docs.edgeimpulse.com/projects/expert-network/driver-drowsiness-detection-with-computer-vision-arduino-nicla-vision.md): Detect closed eyes with FOMO on Arduino Nicla Vision and alert drowsy drivers with a prototype device. - [Gastroscopic Image Processing - OpenMV Cam H7](https://docs.edgeimpulse.com/projects/expert-network/gastroscopic-image-processing-openmv-cam-h7.md): Train an image classification model for gastroscopic images and deploy it to OpenMV Cam H7 and Sony Spresense. - [Pharmaceutical Pill Quality Control and Defect Detection](https://docs.edgeimpulse.com/projects/expert-network/pharmaceutical-pill-defect-detection.md): Detect defects on capsule shells with Edge Impulse FOMO and deploy the model on a Jetson Xavier NX. - [Deter Shoplifting with Computer Vision - Texas Instruments TDA4VM](https://docs.edgeimpulse.com/projects/expert-network/deter-shoplifting-with-computer-vision-ti-tda4vm.md): Train and deploy a bag detector on the TI TDA4VM to demonstrate shoplifting deterrence alerts. - [Smart Factory Prototype - Texas Instruments TDA4VM](https://docs.edgeimpulse.com/projects/expert-network/smart-factory-prototype-ti-tda4vm.md): Deploy an Edge Impulse object detection model on TI SK-TDA4VM hardware for RTSP-based factory equipment monitoring. - [Correct Posture Detection and Enforcement - Texas Instruments TDA4VM](https://docs.edgeimpulse.com/projects/expert-network/ti-tda4vm-posture-detection.md): Train a TDA4VM object detector that monitors posture and controls a relay-driven warning system. - [Visual Anomaly Detection with FOMO-AD - Texas Instruments TDA4VM](https://docs.edgeimpulse.com/projects/expert-network/fomo-ad-ti-tda4vm.md): Train a FOMO-AD visual anomaly detector and run it on a Texas Instruments TDA4VM board. - [Surface Crack Detection and Localization - Texas Instruments TDA4VM](https://docs.edgeimpulse.com/projects/expert-network/surface-crack-detection-ti-tda4vm.md): Run surface crack classification and localization on the TI TDA4VM with TFLite runtime acceleration. - [Surface Crack Detection - Seeed reTerminal](https://docs.edgeimpulse.com/projects/expert-network/surface-crack-detection-seeed-reterminal.md): Deploy a crack localization model on Seeed reTerminal using GAP-CNN training and a desktop app. - [Retail Image Classification - Nvidia Jetson Nano](https://docs.edgeimpulse.com/projects/expert-network/retail-image-classification-nvidia-jetson-nano.md): Train and deploy a Jetson Nano image classification model for recognizing produce in a checkout workflow. - [SiLabs xG24 Plus Arducam - Sorting Objects with Computer Vision and Robotics - Part 1](https://docs.edgeimpulse.com/projects/expert-network/silabs-xg24-card-sorting-and-robotics-1.md): Classify playing cards on a SiLabs xG24 Dev Kit with Arducam and deploy an Edge Impulse image model. - [SiLabs xG24 Plus Arducam - Sorting Objects with Computer Vision and Robotics - Part 2](https://docs.edgeimpulse.com/projects/expert-network/silabs-xg24-card-sorting-and-robotics-2.md): Control a Dobot robot arm with SiLabs xG24 inference results to sort cards and household waste. - [Object Detection and Visualization - Seeed Grove Vision AI Module](https://docs.edgeimpulse.com/projects/expert-network/object-detection-ubidots-seeed-grove-ai.md): Deploy a Grove Vision AI object detector and send inference counts to a Ubidots dashboard. - [Bike Rearview Radar - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/bike-rearview-radar-raspberry-pi.md): Build a Raspberry Pi rearview radar that detects cars behind a bicycle with an object detection model. - [Build a Self-Driving RC Vehicle - Arduino Portenta H7 and Computer Vision](https://docs.edgeimpulse.com/projects/expert-network/arduino-portenta-h7-self-driving-rc-car.md): Build a computer vision RC vehicle with Arduino Portenta H7, object detection training, and vehicle control code. - ['Bring Your Own Model' Image Classifier for Wound Identification](https://docs.edgeimpulse.com/projects/expert-network/arduino-portenta-h7-byom-wound-classification.md): Import a wound classification model into Edge Impulse and deploy it to an Arduino Portenta H7 with a Vision Shield. - [Acute Lymphoblastic Leukemia Classifier - Nvidia Jetson Nano](https://docs.edgeimpulse.com/projects/expert-network/ai-leukemia-classifier-nvidia-jetson-nano.md): Classify leukemia microscopy images on an NVIDIA Jetson Nano using Edge Impulse and the Linux Python SDK. - [Hardhat Detection in Industrial Settings - Alif Ensemble E7](https://docs.edgeimpulse.com/projects/expert-network/helmet-detection-alif-ensemble.md): Train a hardhat detection model and deploy it to an Alif Ensemble E7 kit. - [Motorcycle Helmet Identification and Traffic Light Control - Texas Instruments AM62A](https://docs.edgeimpulse.com/projects/expert-network/motorcycle-helmet-detection-smart-light-ti-am62a.md): Detect motorcycle helmets on Texas Instruments AM62A hardware and control a smart traffic light prototype. - [Import a Pretrained Model with 'Bring Your Own Model' - Texas Instruments AM62A](https://docs.edgeimpulse.com/projects/expert-network/asl-byom-ti-am62a.md): Import an American Sign Language model into Edge Impulse and deploy it on the Texas Instruments AM62A. - [Product Inspection with Visual Anomaly Detection - FOMO-AD - Sony Spresense](https://docs.edgeimpulse.com/projects/expert-network/fomo-ad-product-inspection-spresense.md): Inspect products with FOMO-AD on Sony Spresense and map detected anomalies to hardware feedback. - [Visual Anomaly Detection in Fabric using FOMO-AD - Raspberry Pi 5](https://docs.edgeimpulse.com/projects/expert-network/textile-fabric-anomaly-detection.md): Detect fabric defects on a Raspberry Pi 5 using FOMO-AD and a camera inspection rig. - [Car Detection and Tracking System for Toll Plazas - Raspberry Pi AI Kit](https://docs.edgeimpulse.com/projects/expert-network/vehicle-detection-raspberry-pi-ai-kit.md): Train and convert a toll-plaza vehicle detector for Raspberry Pi 5 and Hailo acceleration. - [Visual Anomaly Detection - Seeed Grove Vision AI Module V2](https://docs.edgeimpulse.com/projects/expert-network/visual-anomaly-detection-seeed-vision-ai-2.md): Train and deploy a Grove Vision AI Module V2 model that detects visual anomalies on a production line. - [Object Counting with FOMO - OpenMV Cam RT1062](https://docs.edgeimpulse.com/projects/expert-network/object-counting-fomo-openmv-rt1062.md): Count nuts on a conveyor belt with FOMO, Edge Impulse, and the OpenMV Cam RT1062. - [Visitor Heatmap with FOMO Object Detection - Jetson Orin Nano](https://docs.edgeimpulse.com/projects/expert-network/visitor-heatmap-jetson-orin-nano.md): Build a Jetson Orin Nano FOMO project that maps visitor positions into a dynamic heatmap. - [Getting Started with the Edge Impulse Nvidia TAO Pipeline - Renesas EK-RA8D1](https://docs.edgeimpulse.com/projects/expert-network/getting-started-nvidia-tao-renesas-ekra8d1.md): Train Nvidia TAO image classification models in Edge Impulse and deploy them to the Renesas EK-RA8D1 kit. - [Smart City Traffic Analysis - NVIDIA TAO + Jetson Orin Nano](https://docs.edgeimpulse.com/projects/expert-network/traffic-analysis-tao-jetson-orin.md): Train an NVIDIA TAO vehicle detector and run traffic counting on a Jetson Orin Nano. - [YOLO-based On-Street Smart Parking Meter - Raspberry Pi 5](https://docs.edgeimpulse.com/projects/expert-network/smart-parking-meter-raspberry-pi.md): Train and deploy a YOLOv5 parking meter model on Raspberry Pi 5 to track vehicles and parking rules. - [LLM-powered Doorbell - ESP32](https://docs.edgeimpulse.com/projects/expert-network/ai-doorbell-esp32.md): Build an ESP32S3 doorbell that combines Edge Impulse face detection with audio, ChatGPT, and Telegram. - [Hand Gestures as Game Controller - Raspberry Pi](https://docs.edgeimpulse.com/projects/expert-network/hand-gesture-controller-raspberry-pi.md): Train a Raspberry Pi hand gesture detector and use it to control Python games. - [Recognizing Punch Cards with LattePanda IOTA and Edge Impulse](https://docs.edgeimpulse.com/projects/expert-network/cv-punchcards-lattepanda.md): Recognize printed punch card patterns with a LattePanda IOTA, webcam, and Edge Impulse image classifier. - [ROS 2 Pick and Place System - Arduino Braccio++ Robotic Arm and Luxonis OAK-D](https://docs.edgeimpulse.com/projects/expert-network/robotic-arm-sorting-arduino-braccio.md): Build a ROS 2 pick-and-place system that combines Edge Impulse object detection with an Arduino Braccio arm. - [Optimize a cloud-based Visual Anomaly Detection Model for Edge Deployments](https://docs.edgeimpulse.com/projects/expert-network/fomo-ad-in-aws.md): Compare visual anomaly detection models for cookie inspection and deploy cloud and web inference workflows. - [Text2Video - Synthetic Data - Nvidia Cosmos Predict](https://docs.edgeimpulse.com/projects/expert-network/nvidia-cosmos-predict2-synthetic-data.md): Generate synthetic training images and videos with NVIDIA Cosmos Predict and prepare labeled data for Edge Impulse Studio. - [Rooftop Ice Detection with Things Network Visualization - Nvidia Omniverse Replicator](https://docs.edgeimpulse.com/projects/expert-network/rooftop-ice-synthetic-data-omniverse.md): Generate synthetic icicle datasets in NVIDIA Omniverse, train an Edge Impulse model, and send LoRaWAN alerts from Arduino Portenta H7. - [Surgery Inventory Object Detection - Synthetic Data - Nvidia Omniverse Replicator](https://docs.edgeimpulse.com/projects/expert-network/surgery-inventory-synthetic-data.md): Generate synthetic surgical tool images with NVIDIA Omniverse Replicator and train a Nicla Vision detector. - [NVIDIA Omniverse - Synthetic Data Generation For Edge Impulse Projects](https://docs.edgeimpulse.com/projects/expert-network/nvidia-omniverse-synthetic-data.md): Create a synthetic fruit dataset with NVIDIA Omniverse Replicator, label it, and deploy an object detector to Jetson Nano. - [Computer Vision Object Counting - Avnet RZBoard V2L](https://docs.edgeimpulse.com/projects/expert-network/avnet-rzboard-object-counting.md): Build a computer vision counter on Avnet RZBoard V2L for tracking objects on a conveyor. - [Gesture Appliances Control with Pose Detection - BrainChip AKD1000](https://docs.edgeimpulse.com/projects/expert-network/gesture-appliances-control-brainchip.md): Use pose detection on a BrainChip Akida kit to control home appliances through Home Assistant. - [Counting for Inspection and Quality Control - Nvidia Jetson Nano (TensorRT)](https://docs.edgeimpulse.com/projects/expert-network/quality-control-jetson-nano.md): Build a Jetson Nano TensorRT object detection application that counts pizza toppings for quality control. - [High-resolution, High-speed Object Counting - Nvidia Jetson Nano (TensorRT)](https://docs.edgeimpulse.com/projects/expert-network/high-speed-counting-jetson-nano.md): Count small objects on a fast conveyor belt with high-resolution FOMO on Nvidia Jetson Nano. - [Occupancy Sensing - SiLabs xG24](https://docs.edgeimpulse.com/projects/expert-network/occupancy-sensing-silabs-xg24.md): Detect room occupancy from xG24 microphone data with Edge Impulse and deploy the model back to the board. - [Smart Appliance Control Using Voice Commands - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/smart-appliance-voice-commands-nordic-thingy53.md): Train a Thingy:53 keyword model for voice-controlled appliances and send commands over Bluetooth to an ESP32 relay. - [Glass Window Break Detection - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/glass-break-detection-nordic-thingy53.md): Train and deploy a Nordic Thingy:53 audio classifier that detects glass breaking events. - [Illegal Logging Detection - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/illegal-logging-detection-nordic-thingy53.md): Train and deploy a Nordic Thingy:53 audio classifier that detects logging truck sounds. - [Illegal Logging Detection - Syntiant TinyML](https://docs.edgeimpulse.com/projects/expert-network/illegal-logging-detection-syntiant-tinyml.md): Build a Syntiant TinyML audio classifier that detects chainsaw noise and sends SMS alerts from remote forest sensors. - [Wearable Cough Sensor and Monitoring - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/wearable-cough-sensor-arduino-nano-33.md): Build a Bluetooth cough monitoring prototype that classifies audio on an Arduino Nano 33 BLE Sense. - [Collect Data for Keyword Spotting - Raspberry Pi Pico](https://docs.edgeimpulse.com/projects/expert-network/collect-data-keyword-spotting-raspberry-pi-pico.md): Collect high-sample-rate keyword spotting audio on a Raspberry Pi Pico and convert serial data into WAV files. - [Voice-Activated LED Strip - Raspberry Pi Pico](https://docs.edgeimpulse.com/projects/expert-network/voice-activated-led-controller-raspberry-pi-pico.md): Control an LED strip with Raspberry Pi Pico voice commands and an Edge Impulse keyword model. - [Snoring Detection on a Smart Phone](https://docs.edgeimpulse.com/projects/expert-network/snoring-detection-on-smartphone.md): Train an audio model for snoring detection and deploy it to a smartphone with Edge Impulse WebAssembly. - [Gunshot Audio Classification - Arduino Nano 33 + Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/gunshot-audio-classification-arduino.md): Build an Arduino-based audio classifier that detects gunshots and sends MQTT notifications. - [AI-Powered Patient Assistance - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/ai-patient-assistance-arduino-nano-33.md): Build an Arduino Nano 33 BLE Sense keyword detector that alerts staff when patients request help. - [Acoustic Pipe Leakage Detection - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/acoustic-pipe-leak-detection-arduino-portenta-h7.md): Detect acoustic pipe leaks with an Arduino Portenta H7, Edge Impulse audio modeling, and firmware deployment. - [Location Identification using Sound - Syntiant TinyML](https://docs.edgeimpulse.com/projects/expert-network/location-identification-using-sound-syntiant-tinyml.md): Train a Syntiant TinyML audio classifier that identifies locations from ambient sound and runs on-device. - [Environmental Noise Classification - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/environmental-noise-classification-nordic-thingy53.md): Classify urban noise on a Nordic Thingy:53 using Edge Impulse audio data collection, training, and deployment. - [Running Faucet Detection - Seeed XIAO Sense + Blues Cellular](https://docs.edgeimpulse.com/projects/expert-network/running-faucet-detection-seeed-xiao-blues-wireless.md): Detect running faucet sounds on a Seeed XIAO nRF52840 Sense and send cellular SMS alerts with Blues Wireless. - [Vandalism Detection via Audio Classification - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/vandalism-detection-audio-classification-arduino-nano-33.md): Train an audio classifier on glass-break sounds and deploy alerts on an Arduino Nano 33 BLE Sense. - [Predictive Maintenance Using Audio Classification - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/predictive-maintenance-using-audio-classification-arduino-nano-33.md): Use audio classification on an Arduino Nano 33 BLE Sense to detect machine anomalies for predictive maintenance. - [Porting an Audio Project from the SiLabs Thunderboard Sense 2 to xG24](https://docs.edgeimpulse.com/projects/expert-network/audio-classification-silabs-xg24.md): Port an audio recognition project from Thunderboard Sense 2 to xG24 and validate live classification. - [Environmental Audio Monitoring Wearable - Syntiant TinyML - Part 1](https://docs.edgeimpulse.com/projects/expert-network/environmental-audio-monitoring-wearable-syntiant-tinyml-part-1.md): Train and deploy a Syntiant TinyML audio classifier that alerts wearable users to vehicle sounds with vibration. - [Environmental Audio Monitoring Wearable - Syntiant TinyML - Part 2](https://docs.edgeimpulse.com/projects/expert-network/environmental-audio-monitoring-wearable-syntiant-tinyml-part-2.md): Assemble a wrist-worn Syntiant TinyML device that vibrates when the environmental audio model detects nearby vehicles. - [Keyword Spotting - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/keyword-spotting-nordic-thingy53.md): Train and deploy a Thingy:53 keyword spotting model that recognizes custom audio commands with Edge Impulse. - [Detecting Worker Accidents with Audio Classification - Syntiant TinyML](https://docs.edgeimpulse.com/projects/expert-network/detecting-worker-accidents-syntiant-tinyml.md): Classify worker accident sounds on a Syntiant TinyML board and trigger safety alerts from detected audio events. - [Snoring Detection with Syntiant NDP120 Neural Decision Processor - Arduino Nicla Voice](https://docs.edgeimpulse.com/projects/expert-network/arduino-nicla-voice-syntiant-snoring-detection.md): Train and deploy a Nicla Voice Syntiant audio model that detects snoring and triggers haptic feedback. - [Recognize Voice Commands with the Particle Photon 2](https://docs.edgeimpulse.com/projects/expert-network/voice-commands-particle-photon-2.md): Train a voice command model and deploy it to a Particle Photon 2 relay controller. - [Voice Controlled Power Plug with Syntiant NDP120 (Nicla Voice)](https://docs.edgeimpulse.com/projects/expert-network/voice-controlled-power-plug-nicla-voice.md): Build an offline Nicla Voice power plug that controls relays from local keyword recognition. - [Determining Compressor State with Audio Classification - Avnet RaSynBoard](https://docs.edgeimpulse.com/projects/expert-network/compressor-audio-classification-rasynboard.md): Classify compressor operating states from audio with an Avnet RaSynBoard and deploy the model for inference. - [Developing a Voice-Activated Product with Edge Impulse's Synthetic Data Pipeline](https://docs.edgeimpulse.com/projects/expert-network/synthetic-data-pipeline-keyword-spotting.md): Generate synthetic speech data, train a keyword spotting model, and build a portable LED device. - [Clawrophyll runs on-Device Wake Word Detection - Arduino UNO Q](https://docs.edgeimpulse.com/projects/expert-network/clawrophyll-on-device-wake-word-detection-using-keyword-spotting-arduino-unoq.md): Train and deploy a wake word model for Clawrophyll on Arduino UNO Q using Edge Impulse. - [Predictive Maintenance - Nordic Thingy:91](https://docs.edgeimpulse.com/projects/expert-network/predictive-maintenance-with-nordic-thingy91.md): Train an anomaly detection model on Nordic Thingy:91 accelerometer data and deploy it for predictive maintenance. - [Brushless DC Motor Anomaly Detection](https://docs.edgeimpulse.com/projects/expert-network/brushless-dc-motor-anomaly-detection.md): Use Edge Impulse to detect brushless DC motor anomalies from ODrive telemetry and classify data with Arduino. - [Industrial Compressor Predictive Maintenance - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/compressor-predictive-maintenance-nordic-thingy53.md): Build a Thingy:53 vibration monitor that detects compressor imbalance and deploys the Edge Impulse model to the device. - [Anticipate Power Outages with Machine Learning - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/anticipate-power-outages-arduino-nano-33.md): Use Arduino Nano 33 BLE Sense temperature and voltage data to train a model that predicts power outage risk. - [Faulty Lithium-Ion Cell Identification in Battery Packs - Seeed Wio Terminal](https://docs.edgeimpulse.com/projects/expert-network/faulty-lithium-ion-cell-identification-battery-packs.md): Identify overheated lithium-ion cells in battery packs using thermal data from a Wio Terminal. - [Weight Scale Predictive Maintenance - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/weight-scale-predictive-maintenance-arduino-nano-33.md): Use IMU data from a rotating pallet wrapper to estimate load weight and detect maintenance issues. - [Fluid Leak Detection With a Flowmeter and AI - Seeed Wio Terminal](https://docs.edgeimpulse.com/projects/expert-network/fluid-leak-detection-with-flowmeter-seeed-wio-terminal.md): Detect no-flow, normal-flow, and leak states from water flow sensor data on a Wio Terminal. - [Pipeline Clog Detection with a Flowmeter and AI - Seeed Wio Terminal](https://docs.edgeimpulse.com/projects/expert-network/pipeline-clog-detection-with-flowmeter-seeed-wio-terminal.md): Classify flowmeter patterns for no flow, normal flow, and clogs, then deploy the model to a Wio Terminal. - [Refrigerator Predictive Maintenance - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/refrigerator-predictive-maintenance-arduino-nano-33.md): Collect refrigerator sensor data and train an anomaly detection model for predictive maintenance on Arduino hardware. - [Motor Pump Predictive Maintenance - Infineon PSoC 6 WiFi-BT Pioneer Kit + CN0549](https://docs.edgeimpulse.com/projects/expert-network/motor-pump-cypress-cn0549.md): Train and deploy a motor pump predictive maintenance model using Infineon PSoC 6 hardware and CN0549 vibration data. - [BrickML Demo Project - 3D Printer Anomaly Detection](https://docs.edgeimpulse.com/projects/expert-network/brickml-3d-printer-anomaly-detection.md): Detect 3D printer anomalies with BrickML sensors, Edge Impulse signal processing, and on-device deployment. - [Condition Monitoring - Syntiant TinyML Board](https://docs.edgeimpulse.com/projects/expert-network/condition-monitoring-syntiant-tinyml.md): Build a TinyML condition monitoring prototype that classifies motor vibration states on a Syntiant TinyML board. - [Predictive Maintenance - Commercial Printer - Sony Spresense + CommonSense](https://docs.edgeimpulse.com/projects/expert-network/predictive-maintenance-sony-spresense-commonsense.md): Monitor printer vibration and air quality with Sony Spresense and CommonSense to detect maintenance issues before downtime. - [Vibration Classification with BrainChip's Akida](https://docs.edgeimpulse.com/projects/expert-network/vibration-classification-with-brainchip-akida.md): Combine Akida classification and anomaly scoring to detect abnormal fan vibrations on edge hardware. - [AI-driven Audio and Thermal HVAC Monitoring - SeeedStudio XIAO ESP32](https://docs.edgeimpulse.com/projects/expert-network/multimodal-hvac-failure-anomaly-detection-esp32.md): Build a multimodal ESP32 HVAC monitor that detects faulty fan sounds and thermal cooling anomalies. - [Edge Link: Industrial PLC Gateway with Edge AI on the Arduino UNO Q](https://docs.edgeimpulse.com/projects/expert-network/industrial-plc-gateway-arduino-unoq.md): Train predictive maintenance models for a simulated PLC production line and deploy them through Edge Link on Arduino UNO Q. - [Arduino x K-Way - Outdoor Activity Tracker](https://docs.edgeimpulse.com/projects/expert-network/arduino-kway-outdoor-activity-tracker.md): Build a K-Way hiking wearable that predicts weather, tracks activity, and logs Nicla Sense ME sensor data. - [Arduino x K-Way - Gesture Recognition for Hiking](https://docs.edgeimpulse.com/projects/expert-network/arduino-kway-gesture-recognition-weather.md): Build a K-Way jacket prototype that recognizes hiking gestures and sends weather and checkpoint updates over Bluetooth. - [Arduino x K-Way - TinyML Fall Detection](https://docs.edgeimpulse.com/projects/expert-network/arduino-kway-fall-detection.md): Detect fall anomalies with a K-Way jacket, Arduino Nicla Sense ME, and Bangle.js smartwatch alert workflow. - [Posture Detection for Worker Safety - SiLabs Thunderboard Sense 2](https://docs.edgeimpulse.com/projects/expert-network/posture-detection-for-worker-safety-silabs-thunderboard-sense-2.md): Train a Thunderboard Sense 2 wearable to classify sitting, lifting, and walking postures for worker safety. - [Hand Gesture Recognition - OpenMV Cam H7](https://docs.edgeimpulse.com/projects/expert-network/hand-gesture-recognition-openmv-cam-h7.md): Train a FOMO hand gesture detector and deploy it to an OpenMV Cam H7. - [Arduin-Row, a TinyML Rowing Machine Coach - Arduino Nicla Sense ME](https://docs.edgeimpulse.com/projects/expert-network/arduin-row-tinyml-rowing-machine-coach-arduino-nicla-sense.md): Build a rowing machine coach with Nicla Sense ME motion inference, anomaly detection, and Arduino IoT Cloud feedback. - [Fall Detection using a Transformer Model - Arduino Giga R1 WiFi](https://docs.edgeimpulse.com/projects/expert-network/fall-detection-with-transformer-model-arduino-giga-r1.md): Train and deploy a Transformer-based fall detection model on an Arduino Giga R1 with accelerometer data. - [Bluetooth Fall Detection - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/bluetooth-fall-detection-arduino-nano-33.md): Detect falls with Arduino Nano 33 BLE Sense motion data and send alerts over Bluetooth to a server. - [Monitor Packages During Transit with AI - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/monitor-packages-during-transit-with-ai-arduino-nano-33.md): Detect package handling events with Arduino Nano 33 BLE Sense accelerometer data and update a Firebase-backed interface. - [Smart Baby Swing - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/smart-baby-swing-arduino-portenta-h7.md): Train an audio classifier on baby cries and use an Arduino Portenta H7 to activate a baby swing. - [Warehouse Shipment Monitoring - SiLabs Thunderboard Sense 2](https://docs.edgeimpulse.com/projects/expert-network/warehouse-shipment-monitoring-silabs-thunderboard-sense-2.md): Monitor fragile package orientation with a Thunderboard Sense 2 accelerometer model and BLE notifications. - [Gesture Recognition - Bangle.js Smartwatch](https://docs.edgeimpulse.com/projects/expert-network/gesture-recognition-banglejs-smartwatch.md): Collect smartwatch motion data, train a gesture model, and deploy it back to a Bangle.js watch. - [Gesture Recognition for Patient Communication - SiLabs Thunderboard Sense 2](https://docs.edgeimpulse.com/projects/expert-network/gesture-recognition-patient-communication-silabs-thunderboard-sense-2.md): Build a Thunderboard Sense 2 wearable that classifies patient hand gestures and reports needs over Bluetooth. - [Hospital Bed Occupancy Detection - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/hospital-bed-occupancy-detection-arduino-nano-33.md): Detect hospital bed occupancy with Arduino Nano 33 BLE Sense motion data and an Edge Impulse classifier. - [Porting a Posture Detection Project from the SiLabs Thunderboard Sense 2 to xG24](https://docs.edgeimpulse.com/projects/expert-network/silabs-xg24-posture-detection.md): Port a wearable posture detection project from Thunderboard Sense 2 to SiLabs xG24 and compare inference performance. - [Porting a Gesture Recognition Project from the SiLabs Thunderboard Sense 2 to xG24](https://docs.edgeimpulse.com/projects/expert-network/silabs-xg24-gesture-recognition.md): Clone a gesture recognition project, test it on Thunderboard Sense 2, and port the firmware to SiLabs xG24. - [Continuous Gait Monitor (Anomaly Detection) - Nordic Thingy:53](https://docs.edgeimpulse.com/projects/expert-network/continuous-gait-monitor-nordic-thingy53.md): Collect gait data with a Thingy:53, train classifier and anomaly detection blocks, and deploy wearable firmware. - [Classifying Exercise Activities on a BangleJS Smartwatch](https://docs.edgeimpulse.com/projects/expert-network/classify-exercise-activities-banglejs-smartwatch.md): Classify exercise activities on a Bangle.js smartwatch with accelerometer data, Edge Impulse training, and Espruino deployment. - [Smart Table Tennis Bat with Live Analytics - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/smart-table-tennis-bat-arduino-nano-33.md): Build a smart table tennis bat that classifies strokes from IMU data and shows live OLED analytics. - [Arduino x K-Way - Environmental Asthma Risk Assessment](https://docs.edgeimpulse.com/projects/expert-network/arduino-kway-environmental-asthma-risk-assessment.md): Train a regression model on weather and asthma visit data for a K-Way smart jacket proof of concept. - [Gas Detection in the Oil and Gas Industry - Nordic Thingy:91](https://docs.edgeimpulse.com/projects/expert-network/gas-detection-nordic-thingy91.md): Build a Nordic Thingy:91 gas monitor that trains and deploys a model for detecting hazardous gas leaks. - [Smart HVAC System with a Sony Spresense](https://docs.edgeimpulse.com/projects/expert-network/sony-spresense-smart-hvac-system.md): Build a Sony Spresense prototype that counts room occupants and adjusts HVAC fan zones. - [Smart HVAC System with an Arduino Nicla Vision](https://docs.edgeimpulse.com/projects/expert-network/arduino-nicla-vision-smart-hvac.md): Use Nicla Vision and FOMO object detection to count occupants and control a scale-model HVAC zone system. - [Indoor CO2 Level Estimation - Arduino Portenta H7](https://docs.edgeimpulse.com/projects/expert-network/indoor-co2-level-estimation-arduino-portenta-h7.md): Collect indoor sensor data and train an Edge Impulse model to estimate CO2 levels with Arduino Portenta H7. - [Harmful Gases Detection - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/harmful-gases-detection-arduino-nano-33.md): Train an Arduino Nano 33 BLE Sense gas classifier using MiCS-4514 sensor readings. - [Fire Detection Using Sensor Fusion and TinyML - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/fire-detection-sensor-fusion-arduino-nano-33.md): Detect fire events with Arduino Nano 33 BLE Sense sensor fusion and send notification alerts. - [AI-Assisted Monitoring of Dairy Manufacturing Conditions - Seeed XIAO ESP32C3](https://docs.edgeimpulse.com/projects/expert-network/dairy-manufacturing-with-ai-seeed-xiao-esp32c3.md): Build a XIAO ESP32C3 device that predicts yogurt consistency from sensor data and reports results in Blynk. - [AI-Assisted Air Quality Monitoring - DFRobot Firebeetle ESP32](https://docs.edgeimpulse.com/projects/expert-network/air-quality-monitoring-firebeetle-esp32.md): Build an AIoT air station with FireBeetle ESP32, air quality sensors, Edge Impulse, and a web dashboard. - [Air Quality Monitoring with Sipeed Longan Nano - RISC-V Gigadevice](https://docs.edgeimpulse.com/projects/expert-network/air-quality-monitoring-sipeed-longan-nano-riscv.md): Build a RISC-V air quality monitor with Sipeed Longan Nano, gas sensors, Edge Impulse, and Arduino deployment. - [Methane Monitoring in Mines - Silabs xG24 Dev Kit](https://docs.edgeimpulse.com/projects/expert-network/methane-monitoring-silabs-xg24.md): Build a Silicon Labs xG24 methane monitor that classifies gas sensor readings with an Edge Impulse model. - [Smart Building Ventilation with Environmental Sensor Fusion](https://docs.edgeimpulse.com/projects/expert-network/smart-building-sensor-fusion.md): Combine sound classification and environmental sensor fusion models to control a smart building window and louver prototype. - [Sensor Data Fusion with Spresense and CommonSense](https://docs.edgeimpulse.com/projects/expert-network/environmental-sensor-fusion-commonsense.md): Fuse environmental sensor data from Spresense and CommonSense boards to classify which room a device occupies. - [Fire Detection Using Sensor Fusion - Arduino Nano 33 BLE Sense](https://docs.edgeimpulse.com/projects/expert-network/fire-detection-sensor-fusion-arduino-nano.md): Build a camera and temperature sensor fusion model for fire detection on Arduino Nano 33 BLE Sense. - [Forest Guard Decentralized Edge-AI LoRa Mesh Network for Forest Surveillance](https://docs.edgeimpulse.com/projects/expert-network/forest-guard-esp32.md): Build a solar LoRa mesh network that detects forest fires and gunshots with edge AI nodes. - [8x8 ToF Gesture Classification - Arduino RP2040 Connect](https://docs.edgeimpulse.com/projects/expert-network/tof-gesture-classification-arduino-rp2040-connect.md): Classify fixed and continuous hand gestures from SparkFun 8x8 ToF sensor data. - [Food Irradiation Dose Detection - DFRobot Beetle ESP32C3](https://docs.edgeimpulse.com/projects/expert-network/food-irradiation-dose-detection-dfrobot-beetle-esp32c3.md): Predict food irradiation dose classes with Beetle ESP32-C3 sensors, a PHP logger, and Edge Impulse. - [Applying EEG Data to Machine Learning, Part 1](https://docs.edgeimpulse.com/projects/expert-network/eeg-data-machine-learning-part-1.md): Control a Pong game with Muse EEG eye blinks, then train an Edge Impulse blink classifier. - [Applying EEG Data to Machine Learning, Part 2](https://docs.edgeimpulse.com/projects/expert-network/eeg-data-machine-learning-part-2.md): Use Muse EEG recordings and Edge Impulse to control a Mind Reader app with movement attempts. - [Applying EEG Data to Machine Learning, Part 3](https://docs.edgeimpulse.com/projects/expert-network/eeg-data-machine-learning-part-3.md): Control a mobile robot with Muse EEG data, Edge Impulse classification, and a Wi-Fi receiver. - [Liquid Classification with TinyML - Seeed Wio Terminal + TDS Sensor](https://docs.edgeimpulse.com/projects/expert-network/liquid-classification-seeed-wio-terminal.md): Classify liquids with a Seeed Wio Terminal, a TDS sensor, and an Edge Impulse model. - [AI-Assisted Pipeline Diagnostics and Inspection with mmWave Radar](https://docs.edgeimpulse.com/projects/expert-network/ai-pipeline-inspection-mmwave-radar.md): Build a Nicla Vision and mmWave radar system that diagnoses pipeline defects and logs results in a web application. - [Soil Quality Detection Using AI and LoRaWAN - Seeed Sensecap A1101](https://docs.edgeimpulse.com/projects/expert-network/soil-quality-detection-lorawan-seeed-sensecap-a1101.md): Detect fertilizer contamination with SenseCAP A1101, FOMO object detection, LoRaWAN, and WhatsApp alerts. - [Smart Diaper Prototype - Arduino Nicla Sense ME](https://docs.edgeimpulse.com/projects/expert-network/smart-diaper-arduino-nicla-sense-me.md): Use Nicla Sense ME gas and air quality readings to detect a soiled diaper and report status by Bluetooth. - [DIY Smart Glove with Flex Sensors](https://docs.edgeimpulse.com/projects/expert-network/flex-sensors-hci.md): Build a low-cost smart glove with DIY flex sensors and classify gestures with Edge Impulse. - [EdgeML Energy Monitoring - Particle Photon 2](https://docs.edgeimpulse.com/projects/expert-network/energy-monitoring-particle-photon-2.md): Build an edge ML energy monitor with Particle Photon 2 to identify appliances from shared AC current signatures. - [Wearable for Monitoring Worker Stress using HR/HRV DSP Block - Arduino Portenta](https://docs.edgeimpulse.com/projects/expert-network/ecg-hrv-block-arduino-portenta.md): Use the HR/HRV processing block with Arduino Portenta to classify ECG-based worker stress signals. - [Azure Machine Learning with Kubernetes Compute and Edge Impulse](https://docs.edgeimpulse.com/projects/expert-network/azure-machine-learning-kubernetes-compute-ei.md): Connect Azure Machine Learning Kubernetes compute with Edge Impulse using a Raspberry Pi audio example. - [ROS2 + Edge Impulse, Part 1: Pub/Sub Node in Python](https://docs.edgeimpulse.com/projects/expert-network/ros2-part1-pubsub-node.md): Build a reusable ROS2 Python node that subscribes to sensor data and publishes Edge Impulse inference results. - [ROS2 + Edge Impulse, Part 2: MicroROS](https://docs.edgeimpulse.com/projects/expert-network/ros2-part2-microros.md): Run an Edge Impulse vision model on an Arduino Portenta H7 and publish MicroROS inference messages to ROS2. - [Using Hugging Face Datasets in Edge Impulse](https://docs.edgeimpulse.com/projects/expert-network/huggingface-dataset-with-edge-impulse.md): Upload a Hugging Face beans dataset with the Edge Impulse CLI and deploy the trained model. - [Using Hugging Face Image Classification Datasets with Edge Impulse](https://docs.edgeimpulse.com/projects/expert-network/hugging-face-image-classification-dataset.md): Import Hugging Face image classification data into Edge Impulse and train a model. - [Offline SLMs for Edge AI Development, Part 1: Direct Inference with a Qwen LoRA Adapter](https://docs.edgeimpulse.com/projects/expert-network/integrating-slms-on-linux.md): Run a fine-tuned Qwen2.5-Coder LoRA adapter locally on a Raspberry Pi, Rubik Pi 3, or NVIDIA Jetson Orin to answer Edge Impulse docs and API questions without an internet connection. - [Offline SLMs for Edge AI Development - Part 2: Retrieval-Augmented Generation with FAISS](https://docs.edgeimpulse.com/projects/expert-network/rag-docs-assistant-faiss-qwen.md): Pair a FAISS retrieval index with a Qwen2.5-Coder LoRA adapter to answer grounded offline questions about Edge Impulse documentation. - [Offline SLMs for Edge AI Development, Part 3: Agentic Coding with llama.cpp and OpenCode](https://docs.edgeimpulse.com/projects/expert-network/opencode-offline-coding-assistant.md): Set up an offline agentic coding environment on a Raspberry Pi, Rubik Pi 3, or NVIDIA Jetson using OpenCode, llama.cpp, and two fine-tuned Qwen2.5-Coder GGUF models. - [Edge Intelligence Pipeline: LLM-Enhanced Interpretation of Embedded ML Inference](https://docs.edgeimpulse.com/projects/expert-network/llm-enhanced-image-inference-pipeline.md): Combine Raspberry Pi image classification with an LLM pipeline that turns embedded inference results into contextual explanations. - [Community Guide - Using Edge Impulse with Nvidia DeepStream](https://docs.edgeimpulse.com/projects/expert-network/nvidia-deepstream-community-guide.md): Export Edge Impulse models for NVIDIA DeepStream and configure TensorRT inference pipelines for vision projects. - [Edge Impulse API Usage Sample Application - Jetson Nano Trainer](https://docs.edgeimpulse.com/projects/expert-network/api-sample-application-jetson-nano.md): Use the Edge Impulse API from a Jetson Nano Python trainer to create, train, test, and run a vision model. - [MLOps with Edge Impulse and Azure IoT Edge](https://docs.edgeimpulse.com/projects/expert-network/mlops-azure-iot-edge.md): Connect Edge Impulse with Azure IoT Edge to collect data, deploy models, and retrain an edge workflow. - [A Federated Approach to Train and Deploy Machine Learning Models](https://docs.edgeimpulse.com/projects/expert-network/federated-learning-raspberry-pi.md): Train a federated image classification model with Flower and deploy the global model through Edge Impulse. - [DIY Model Weight Update for Continuous AI Deployments](https://docs.edgeimpulse.com/projects/expert-network/diy-model-ota.md): Patch an Edge Impulse Arduino library to update model weights over the air without reflashing firmware. - [Automate the CI/CD Pipeline of your Models with Edge Impulse and GitHub Actions](https://docs.edgeimpulse.com/projects/expert-network/cicd-with-github-actions-edge-impulse.md): Automate Edge Impulse model workflows with GitHub Actions for CI/CD across data, training, deployment, and releases. - [Deploying Edge Impulse Models on ZEDEDA Cloud Devices](https://docs.edgeimpulse.com/projects/expert-network/zededa-model-monitoring.md): Package an Edge Impulse object detection model as a container and deploy it to ZEDEDA-managed edge devices. - [Building a Voice-Activated Assistant on Your Smartphone: A Step-by-Step Guide](https://docs.edgeimpulse.com/projects/expert-network/android-keyword-spotting.md): Build a custom Android wake-word detector with Edge Impulse audio models, TensorFlow Lite, and native C++ integration. ## Datasets - [Datasets](https://docs.edgeimpulse.com/datasets.md): Browse public Edge Impulse datasets for image, audio, and time-series machine learning experiments. - [Fire extinguisher safety pin](https://docs.edgeimpulse.com/datasets/image/fire-extinguisher-safety-pin.md): A labeled image dataset of fire extinguisher safety pins for training image classification and visual anomaly detection models. - [Microscope](https://docs.edgeimpulse.com/datasets/image/microscope.md): A labeled image dataset captured through a smartphone microscope for training image classification models. - [Bottles rack](https://docs.edgeimpulse.com/datasets/image/bottles-rack.md): A labeled image dataset of bottle racks for training object detection models to identify filled and empty bottle slots. - [Cans on conveyor belt](https://docs.edgeimpulse.com/datasets/image/cans-on-conveyor-belt.md): A synthetically generated image dataset of cans on a conveyor belt for training object detection models using NVIDIA Omniverse. - [Cubes on conveyor belt (colors)](https://docs.edgeimpulse.com/datasets/image/cubes-on-conveyor-belt-colors.md): A labeled image dataset of colored cubes on a conveyor belt, used to design and demonstrate the FOMO object detection architecture. - [Cubes on conveyor belt (self attention)](https://docs.edgeimpulse.com/datasets/image/cubes-on-conveyor-belt-self-attention.md): A labeled image dataset of conveyor belt cubes for spatial-aware object detection using a micro transformer with FOMO self-attention. - [Dice](https://docs.edgeimpulse.com/datasets/image/dice.md): A labeled image dataset of dice on a white background for training object detection models to count and locate objects in a frame. - [Dice colors](https://docs.edgeimpulse.com/datasets/image/dice-colors.md): A labeled image dataset of colored dice on a white background for training object detection models to classify dice by color. - [Capsule](https://docs.edgeimpulse.com/datasets/image/capsule.md): A labeled image dataset of pharmaceutical capsules for training visual anomaly detection models to identify defects. - [DHT11](https://docs.edgeimpulse.com/datasets/image/dht11.md): A labeled image dataset of DHT11 sensors for training visual anomaly detection models to identify damaged or mis-wired components. - [Fire extinguisher head thread](https://docs.edgeimpulse.com/datasets/image/fire-extinguisher-head-thread.md): A labeled image dataset of fire extinguisher head threads for visual anomaly detection in industrial quality inspection. - [Flat washers](https://docs.edgeimpulse.com/datasets/image/flat-washers.md): A labeled image dataset of flat washers on varying backgrounds for training visual anomaly detection models. - [Thermostatic valves](https://docs.edgeimpulse.com/datasets/image/thermostatic-valves.md): A labeled image dataset of thermostatic valves for training visual anomaly detection models in industrial quality inspection. - [Vial tubes](https://docs.edgeimpulse.com/datasets/image/vial-tubes.md): A labeled image dataset for training visual regression models to predict liquid fill levels in vial tubes. - [Faucet vs noise](https://docs.edgeimpulse.com/datasets/audio/faucet-vs-noise.md): A labeled audio dataset of faucet water sounds vs. background noise for training audio classification models on embedded devices. - [Keyword Spotting](https://docs.edgeimpulse.com/datasets/audio/keyword-spotting.md): A labeled audio dataset for training keyword spotting models to detect the 'Hello World' phrase on embedded devices. - [Coffee machine stages (vibration)](https://docs.edgeimpulse.com/datasets/time-series/coffee-machine-stages-vibration.md): A vibration time-series dataset of coffee machine operation stages for training motion and vibration classification models. - [Continuous motion recognition](https://docs.edgeimpulse.com/datasets/time-series/continuous-motion-recognition.md): A labeled accelerometer dataset of continuous gestures (idle, snake, up-down, wave) for training motion recognition models. - [Coffee machine stages (sensor fusion)](https://docs.edgeimpulse.com/datasets/time-series/coffee-machine-stages-sensor-fusion.md): A multi-sensor time-series dataset of coffee machine operation stages for demonstrating sensor fusion classification with neural network embeddings. ## Optional - [Forum](https://forum.edgeimpulse.com): Community and support resources. - [Discord](https://discord.gg/edgeimpulse): Community and support resources. - [GitHub](https://github.com/edgeimpulse): Community and support resources.