Disentangling the modes of variation in unlabelled data
- Submitting institution
-
Middlesex University
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 774
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2017.2783940
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 2682
- Volume
- 40
- Issue
- 11
- ISSN
- 0162-8828
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2017
- URL
-
http://eprints.mdx.ac.uk/23771/
- 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)
-
-
- Citation count
- 4
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Statistical methods are of paramount importance in discovering the modes of variation in visual data. This paper proposes a novel general multilinear matrix decomposition method that discovers the multilinear structure of possibly incomplete sets of visual data in unsupervised setting. This is significant as the extensions of the method are able to handle noisy data and to exploit available geometric or label information for some modes of variations, leading to significant contributions to several computer vision tasks, including Shape from Shading, expression transfer, and estimation of surface normals from images captured in the wild.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -