Wing Loss for Robust Facial Landmark Localisation with Convolutional Neural Networks
- Submitting institution
-
The University of Surrey
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 9018801_4
- Type
- E - Conference contribution
- DOI
-
10.1109/CVPR.2018.00238
- Title of conference / published proceedings
- 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
- First page
- 0
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- -
- Year of publication
- 2018
- 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
-
-
- Research group(s)
-
-
- Citation count
- 40
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- The key contribution of this work is to propose a new loss function for regression-based facial landmark detection with CNNs. To the best of our knowledge, this is the first such work in the community. The new loss function improves the performance of different CNN architectures in facial landmark detection as compared with traditionally used L2, L1 and Smooth L1 loss functions. We further improved this loss and found that the new loss function can promote the performance of light-weight networks more significantly, which generates positive impacts on many practical applications on portable devices.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -