Mesh Saliency via Weakly Supervised Classification-for-Saliency CNN
- Submitting institution
-
University of Brighton
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 7153712
- Type
- D - Journal article
- DOI
-
10.1109/TVCG.2019.2928794
- Title of journal
- IEEE Transactions on Visualization and Computer Graphics
- Article number
- -
- First page
- 151
- Volume
- 27
- Issue
- 1
- ISSN
- 1077-2626
- Open access status
- Compliant
- Month of publication
- July
- 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
-
2
- Research group(s)
-
-
- Citation count
- 2
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Detecting saliency in 3D models using machine learning usually requires a large amount of data, or annotated 3D models, as ground truth for training the neural networks. This paper is significant because it proposes a novel Classification-for-Saliency CNN which relies on transferring knowledge from 3D object classification to mesh saliency and does not require any saliency ground truth, but only the class membership of meshes. The research has been used as a baseline for other approaches researching on detecting saliency via machine learning, for example, in industrial modelling applications (Arvanitis et al., IEEE Transactions on Industrial Informatics, 2021).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -