DensePose: Dense Human Pose Estimation in the Wild.
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 16245
- Type
- E - Conference contribution
- DOI
-
10.1109/CVPR.2018.00762
- Title of conference / published proceedings
- CVPR
- First page
- 7297
- Volume
- -
- Issue
- -
- ISSN
- 2575-7075
- Open access status
- Not compliant
- Month of publication
- December
- 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
-
2
- Research group(s)
-
-
- Citation count
- 94
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This work introduced and tackled the task of densely regressing human surface coordinates from an RGB image, which involved gathering and training with a large dataset of densely annotated humans. This work was cited hundreds of times in 2 years by research labs in MIT, Stanford, Berkeley, CMU, Google, Amazon, Adobe, among others. The DensePose videos gathered more than 500K views, the github repo has 5.8K stars, while Mark Zuckerberg included DensePose as one among four samples of Facebook’s open research when testifying in congress. This work led me to co-found Ariel AI, a UK startup for 3D human reconstruction.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -