Deep Affect Prediction in-the-Wild: Aff-Wild Database and Challenge, Deep Architectures, and Beyond
- Submitting institution
-
Goldsmiths' College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 3078
- Type
- D - Journal article
- DOI
-
10.1007/s11263-019-01158-4
- Title of journal
- International Journal of Computer Vision
- Article number
- -
- First page
- 907
- Volume
- 127
- Issue
- 6-7
- ISSN
- 0920-5691
- Open access status
- Compliant
- Month of publication
- June
- Year of publication
- 2019
- URL
-
http://research.gold.ac.uk/id/eprint/26111/
- 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
-
7
- Research group(s)
-
-
- Citation count
- 15
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This work introduces a new database for benchmarking the performance of algorithms on the problem of audio-visual analysis of continuous affective behaviour. The dataset was part of the first ever challenge that focused on estimation of continuous affect in uncontrolled settings, held in conjunction with CVPR 2017. Beyond the dataset, novel deep architectures are proposed that produce state-of-the-art results on this and other benchmarks – by using the newly collected dataset as a prior for training. This work has both pushed the boundaries of the state of the art and brought more attention to this challenging problem in the community.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -