Frame from the EPIC-KITCHENS-100 dataset, University of Bristol, CC BY-NC 4.0.
CC BY-NC 4.0
The publisher's licence restricts use to research and other non-commercial purposes.
git clone https://github.com/epic-kitchens/epic-kitchens-download-scripts && cd epic-kitchens-download-scripts && python epic_downloader.py --videosFiles are hosted openly at https://data.bris.ac.uk/datasets/3h91syskeag572hl6tvuovwv4d/ and can be pulled per participant, per video, or per modality (--videos, --rgb-frames, --flow-frames, --extension-only, --participants). Download speed from data.bris can be very slow outside the UK; Academic Torrents mirrors exist for both original (1.1 TB) and extended (740 GB) sequences. Annotations are on GitHub. CC BY-NC 4.0; commercial licence by email to uob-epic-kitchens@bristol.ac.uk.
Specification, as published
- Hours (stated)
- 100 h
- Episodes (stated)
- 700 episodes
- Scale
- 100 hours, 20M frames, 90K actions in 700 variable-length videos
- Task
- Cooking & food prep, Dishwashing, Cleaning, Tidying & put-away
- Environment
- Indoor – home
- Modality
- RGB, Audio, IMU, Language, Hand pose
- Capture
- Egocentric
- In frame
- Person hands
- Captured in
- Multiple regions
- Embodiment
- human
- Frame rate
- 50 fps
- Resolution
- 1920x1080
- Formats
- MP4, Images (JPEG/PNG), Pickle, Other
- Size
- about 1.84 TB zipped (Original Sequences 1.1 TB plus Extended Sequences 740 GB)
- Hosted on
- data.bris.ac.uk
Figures are the publisher's own. Kinetic Blocks has not measured this dataset.
About this dataset
Participants recorded every visit to their own kitchen with a head-mounted camera, producing 100 hours and 20 million frames of unscripted cooking, cleaning and tidying across 45 kitchens in 4 cities in Europe and North America. Videos are 1920x1080 at 50 fps with audio; the release adds RGB and optical-flow frames, gyroscope and accelerometer data, 90K narrated action segments with 97 verb and 300 noun classes, and hand-object and mask annotations from sibling projects. Files are served openly from the University of Bristol data repository and Academic Torrents; the licence is CC BY-NC 4.0 with commercial licences available by email.
Publisher's own task names: unscripted kitchen activities, action recognition, action detection, action anticipation, multi-instance retrieval, unsupervised domain adaptation, video object segmentation, 97 verb classes, 300 noun classes.
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.