EPIC-Fusion : Audio-Visual Temporal Binding for Egocentric Action Recognition
- Submitting institution
-
University of Bristol
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 219667126
- Type
- E - Conference contribution
- DOI
-
10.1109/ICCV.2019.00559
- Title of conference / published proceedings
- 2019 IEEE/CVF International Conference on Computer Vision (ICCV)
- First page
- 5491
- Volume
- -
- Issue
- -
- ISSN
- 2380-7504
- Open access status
- Compliant
- Month of publication
- November
- Year of publication
- 2019
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
3
- Research group(s)
-
C - Visual Information Lab
- Citation count
- 5
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- The first work to explore audio-visual fusion for egocentric actions, making use of the large-scale dataset EPIC-KITCHENS (http://epic-kitchens.github.io). This collaboration with the Oxford University provided background for EPSRC Programme Grant: Visual AI (EP/T028572/1, 2020-2025). Work is inspired from temporal binding in experimental psychology, and also provided background to Damen’s EPSRC Fellowship UMPIRE (EP/T004991/1, 2020-2025). Public code and model checkpoints [https://github.com/ekazakos/temporal-binding-network] frequently forked (70 stars). The method was second ranked in the 2019 Action Recognition challenge on the EPIC-KITCHENS dataset. Features from this model were used for follow-up work on domain adaptation [Munro and Damen, CVPR 2020].
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -