Dense 3D object reconstruction from a single depth view
- Submitting institution
-
The University of Warwick
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 6129
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2018.2868195
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 2820
- Volume
- 41
- Issue
- 12
- ISSN
- 0162-8828
- Open access status
- Compliant
- Month of publication
- December
- 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)
-
I - Artificial Intelligence and Human-Centred Computing
- Citation count
- 12
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Published in the leading journal on pattern recognition and machine intelligence, this work introduced a generative adversarial learning framework 3D-RecGAN, which was the first to reconstruct the full 3D structure of an object at high resolution from a single arbitrary-depth view. It achieved new state-of-the-art performance, and released to the community an open dataset with high-quality ground truth collected in real-world scenarios. This research has been taken up and competed with in a series of subsequent works (e.g. Tombari, Google, ICCV 2019; Berenson, Michigan, ISRR 2019). It has also led to talk invitations from top institutions (NTU, Sun Yat-sen, CSIRO).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -