Frame from the HD-EPIC dataset, University of Bristol, CC BY 4.0.
CC BY 4.0
The publisher's licence permits commercial use and redistribution, with attribution.
wget -r -np -nH --cut-dirs=2 https://data.bris.ac.uk/datasets/3cqb5b81wk2dc2379fx1mrxh47/Open deposit (2.3 TiB) on the University of Bristol data repository, organised per participant P01 to P09 with separate folders for video, VRS, audio, hand masks, SLAM and gaze, and digital twins. data.bris pages time out intermittently and downloads can be slow outside the UK. The wget command is a generic recursive fetch of the DOI folder and has not been verified end to end; check the record page for an official download script. Annotations and some intermediate data (Dropbox, SharePoint links) are in the GitHub repo.
Specification, as published
- Hours (stated)
- 41 h
- Scale
- 41 hours of unscripted multi-day recordings
- Task
- Cooking & food prep, Dishwashing, Cleaning, Tidying & put-away
- Environment
- Indoor – home
- Modality
- RGB, Audio, Eye gaze, IMU, Point cloud, Language, Hand pose
- Capture
- Egocentric
- In frame
- Person hands
- Captured in
- Europe
- Embodiment
- human
- Frame rate
- 30 fps
- Resolution
- 1408x1408
- Formats
- MP4, HDF5, JSON, Pickle, Other
- Size
- 2.3 TiB
- Hosted on
- data.bris.ac.uk
Figures are the publisher's own. Kinetic Blocks has not measured this dataset.
About this dataset
Nine participants wore Project Aria glasses in their own kitchens for at least three consecutive days, giving 41 hours and 4.4 million frames of unscripted cooking with 1408x1408 RGB at 30 fps, 7-microphone audio, eye gaze and SLAM camera poses. Every kitchen was reconstructed as a 3D digital twin, and the footage carries 69 recipes, 59.4K actions, 50.9K audio events, 7.7M hand masks, 19.9K object tracks and a 26.6K-question VQA benchmark. The 2.3 TiB deposit on the University of Bristol data repository is open and marked CC BY 4.0; annotations are on GitHub.
Publisher's own task names: 69 recipes, recipe recognition, action recognition, object movement, audio event detection, gaze estimation, VQA benchmark (7 categories), 3D digital twins.
What Kinetic Blocks did, and did not do
Indexed and linked from the publisher. Not hosted, verified or graded by Kinetic Blocks. The publisher's terms govern. The licence shown is the one stated on the publisher's page, linked above. Nothing on this page is a claim by Kinetic Blocks.
Want the verified ones too?
Request access to the marketplace, where every paid listing is verified against its files and graded before it goes on sale.