Gauss-newton deformable part models for face alignment in-the-wild
- Submitting institution
-
Imperial College of Science, Technology and Medicine
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2226
- Type
- E - Conference contribution
- DOI
-
10.1109/CVPR.2014.239
- Title of conference / published proceedings
- 2014 IEEE conference on computer vision and pattern recognition (CVPR)
- First page
- 1851
- Volume
- -
- Issue
- -
- ISSN
- 1063-6919
- Open access status
- Out of scope for open access requirements
- Month of publication
- September
- Year of publication
- 2014
- URL
-
-
- Supplementary information
-
10.1109/CVPR.2014.239
- 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
-
1
- Research group(s)
-
-
- Citation count
- 127
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- The paper presents the first robust and computational efficient part-based Active Appearance Model algorithm. The algorithm was applied for accurate tracking of elderly faces by webcams for tele-presence robots (EC H2020 TERESA project; http://teresaproject.eu) and for subsequent detection of markers of depression, based on the tracked dynamics of certain facial landmarks (ITV News 17/Feb/17; Channel 4 "Old People's Home for 4 Years Old", July 2017). It resulted in a lot of interest by media outlets (e.g. CBS 60 Minutes (USA) 2017, Arirang Special (Korea) 2017) and led to a keynote at PervasiveHealth 2017 (http://archive.pervasivehealth.org/2017/show/keynotes). CVPR 2014 acceptance rate: 29.8%/540.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
