Eidon Tracker POV: Egocentric Household Video with IMU Arm Tracking
- Provider
- Eidon AI (Solidic Labs Inc)
- License
- CC-BY-4.0 (source)
- Size
- 13,455 files
- Format
- MP4, Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/eidon-ai/tracker-pov)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Robotics
- Source
- https://huggingface.co/eidon-ai
About Eidon Tracker POV: Egocentric Household Video with IMU Arm Tracking
Eidon Tracker POV is a final release of 1,274 hours of egocentric video recorded while contributors did ordinary household work in their own homes, such as laundry, cleaning, dishes and cooking. Contributors wore a head-mounted camera and a seven-sensor IMU harness, so each recording pairs first-person video with 24 Hz orientation data for both hands, both forearms, both upper arms and the chest. The card reports 13,451 recordings from 27 contributors; video is MP4 from 1080p to 4K, mostly at 30 fps. This repo holds the video under recordings/data/ (in numbered folders of 1,000 files) plus recordings/metadata.parquet, one row per recording with duration, task type, timestamp and contributor id, and users.parquet. The matching IMU streams are in the separate eidon-ai/tracker-pov-imu repo (about 9.5 GB, joined on recording_id). Measured 2026-11-14 at about 9 TB. Read metadata with pandas and video with PyAV, decord or OpenCV. Useful for learning manipulation from human video, activity recognition and motion-to-video modelling. Published under CC BY 4.0, with attribution required.
How to load Eidon Tracker POV: Egocentric Household Video with IMU Arm Tracking in Python
from datasets import load_dataset
ds = load_dataset("eidon-ai/tracker-pov")
print(ds)Data structure
File types in sampled files: MP4, Parquet
Files
Example file: recordings/metadata.parquet
Analyze this dataset without downloading it
On Scigantic, Eidon Tracker POV: Egocentric Household Video with IMU Arm Tracking opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Robotics datasets
- PhysicalAI Robotics Manipulation Objects Kitchen MJCF
- IsaacLab-Arena Unitree G1 Humanoid Apple Pick-and-Place: 200 XR-Teleoperated Simulation Demonstrations (NVIDIA)
- GrandTour: Legged Robot Multi-Sensor Dataset (ETH Zurich)
- PhysicalAI Traffic Anomaly Reasoning
- PhysicalAI Robotics Manipulation Kitchen Demos
- Dexora Real-World Dataset: High-DoF Bimanual Dexterous Teleoperation
- PhysicalAI Robotics GR 00 T GR 1
- ArtVIP: Articulated Digital Assets for Robot Learning in Isaac Sim