3DPalsyNet : A Facial Palsy Grading and Motion Recognition Framework using Fully 3D Convolutional Neural Networks
- Submitting institution
-
University of Northumbria at Newcastle
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 26077980
- Type
- D - Journal article
- DOI
-
10.1109/ACCESS.2019.2937285
- Title of journal
- IEEE Access
- Article number
- -
- First page
- 121655
- Volume
- 7
- Issue
- -
- ISSN
- 2169-3536
- Open access status
- Compliant
- Month of publication
- August
- Year of publication
- 2019
- 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
-
4
- Research group(s)
-
C - Digital Learning Laboratory (DLL)
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper proposes a new end-to-end framework for facial palsy grading using a 3D Convolutional Neural Network architecture with a ResNet backbone for the prediction of these dynamic tasks. The evaluation has shown an attractive level of classification accuracy of up to 86% including further improvement in spatio-temporal feature learning when compared to softmax loss alone. The work was carried out in collaboration with Prof Chang-Tsun Li from Deakin University, Australia. A further exploitation jointly with industry partners is underway.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
