Learning to transfer: transferring latent task structures and its application to person-specific facial action unit detection
- Submitting institution
-
University of Nottingham, The
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1332810
- Type
- E - Conference contribution
- DOI
-
10.1109/ICCV.2015.430
- Title of conference / published proceedings
- 2015 IEEE International Conference on Computer Vision (ICCV 2015)
- First page
- 3774
- Volume
- 2015-Dec
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- December
- Year of publication
- 2015
- 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
-
2
- Research group(s)
-
-
- Citation count
- 17
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Learning to transfer latent task structures allows a machine learning system to improve performance of facial expression analysis on previously unseen people, by focusing on the similarities in expressive behaviour between a new person and those in a database. This is significant because obtaining person-specific facial expression data requires rare skills, is time-consuming and expensive. This paper has led to a renewed focus on person-specific facial expression recognition using weak or semi-supervised methods (>8papers on this topic). In addition, it has inspired many multi-task solutions, with at least 10 papers building on the paper’s findings.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
