Dynamic Probabilistic CCA for Analysis of Affective Behaviour and Fusion of Continuous Annotations
- Submitting institution
-
Goldsmiths' College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2013
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2014.16
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 1299
- Volume
- 36
- Issue
- 7
- ISSN
- 0162-8828
- Open access status
- Out of scope for open access requirements
- Month of publication
- July
- Year of publication
- 2014
- URL
-
http://research.gold.ac.uk/id/eprint/17311/
- 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
- 33
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Obtaining reliable annotations for the massive amounts of data available, that are in turn necessary to train machine learning algorithms is a tedious, laborious, error-prone and expensive task. The authors present a novel probabilistic approach that enables tackling major problems in obtaining reliable data annotations in continuous time and space – such as temporal discrepancies and annotator reliability. This work has been subsequently used for a multitude of relevant applications, including affect-aware robotics (FROG, FP7/2007-2013), and pain estimation (EmoPain - EPSRC EP/H017178/1), and has brought attention to issues arising from annotated data leading to increased interest by other researchers.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
