How far are we from solving the 2D & 3D face alignment problem? (and a dataset of 230,000 3D facial landmarks)
- Submitting institution
-
University of Nottingham, The
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1330511
- Type
- E - Conference contribution
- DOI
-
10.1109/ICCV.2017.116
- Title of conference / published proceedings
- 2017 IEEE International Conference on Computer Vision (ICCV 2017)
- First page
- 1021
- Volume
- 2017-October
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- December
- Year of publication
- 2017
- 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
-
1
- Research group(s)
-
-
- Citation count
- 125
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Face alignment methods are employed regularly in face analysis systems. This paper not only presents a state of the art method for face alignment, and shows that it is capable of better quality detections than human annotators, but also explores how far we are from actually solving this problem completely. Finally, the paper offers a very large synthetically augmented dataset of facial point annotations for corresponding images. The paper was published at a high-ranking conference with an average acceptance rate of 26%.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
