HandMap: robust hand pose estimation via intermediate dense guidance map supervision
- Submitting institution
-
Cardiff University / Prifysgol Caerdydd
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 97032654
- Type
- E - Conference contribution
- DOI
-
10.1007/978-3-030-01270-0_15
- Title of conference / published proceedings
- Lecture Notes in Computer Science
- First page
- 246
- Volume
- 0
- Issue
- -
- ISSN
- 0302-9743
- Open access status
- Technical exception
- Month of publication
- October
- Year of publication
- 2018
- URL
-
https://doi.org/10.1007/978-3-030-01270-0_15
- 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
-
3
- Research group(s)
-
C - Cybersecurity, privacy and human centred computing
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- We apply dense guidance maps (e.g., vectors from each pixel pointing towards each finger joint position) as a residual module from the input image in a supervised learning framework to improve the accuracy of hand pose estimation. We further optimize our solution by approximating the distance from each pixel to each joint via propagation using a fast-marching method. Our work can be easily inserted into existing learning pipelines, achieving the best performance compared to state-of-the-art methods. Our work was presented at the ECCV international conference in 2018 (pp. 237-253).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -