Deep Sign: Enabling Robust Statistical Continuous Sign Language Recognition via Hybrid CNN-HMMs
- Submitting institution
-
The University of Surrey
- Unit of assessment
- 12 - Engineering
- Output identifier
- 9003934_1
- Type
- D - Journal article
- DOI
-
10.1007/s11263-018-1121-3
- Title of journal
- International Journal of Computer Vision
- Article number
- -
- First page
- 1311
- Volume
- 126
- Issue
- 12
- ISSN
- 0920-5691
- Open access status
- Compliant
- 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)
-
-
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Selected as a top paper published at BMVC2016 this invited, extended version presents a hybrid CNN-HMM architecture for recognition of spatio-temporal sequences combining strong discriminative neural networks with the sequence modelling capabilities of Bayesian networks. The paper demonstrated how the power of deep learning and Bayesian models can be combined to provide a step change in performance. Applied to sign language recognition it achieves a significant error reduction over multiple datasets.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -