A novel performance evaluation methodology for single-target trackers
- Submitting institution
-
The University of Birmingham
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 30659921
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2016.2516982
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 2137
- Volume
- 38
- Issue
- 11
- ISSN
- 0162-8828
- Open access status
- Out of scope for open access requirements
- Month of publication
- January
- Year of publication
- 2016
- 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
-
8
- Research group(s)
-
-
- Citation count
- 203
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- The paper proposes a novel and comprehensive performance evaluation methodology for single-target visual trackers. A significant novelty is the use and first-of-its-kind analysis of re-initialisations at tracking failures as well as the novel ranking methodology. The methodology was tested on 38 trackers, making it the largest benchmark at the time of publication. The proposed methodology has become a standard for performance evaluation of visual trackers and forms the basis for visual tracking challenges (VOT) which take place yearly in conjunction with the major computer vision conferences (ICCV’13,’15,’17,’19; ECCV’14,’16,’18). The VOT2019 challenge attracted participation of 57 visual trackers (http://www.votchallenge.net/vot2019/).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -