# Edge Impulse Documentation - Datasets

> Reusable image, audio, and time-series datasets for prototyping.

This is the "Datasets" section index of the Edge Impulse documentation, following https://llmstxt.org/. It lists the pages in this section only. For other sections see the root index at https://docs.edgeimpulse.com/llms.txt; for the complete expanded index see https://docs.edgeimpulse.com/llms-full.txt.

Links point to the markdown-rendered version of each page (`.md`), so agents retrieve clean content without navigation, ads, or scripts. Every link includes a short note describing the page.

## Intents

Intent-bearing retrieval prompts. Match a user request to the closest intent below, then retrieve the linked pages before answering.

### 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)

## Datasets

- [Datasets](https://docs.edgeimpulse.com/datasets.md)
- [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

- [Documentation index (all sections)](https://docs.edgeimpulse.com/llms.txt): Root llms.txt router linking every documentation section.