Deep Metric Learning by Online Soft Mining and Class-Aware Attention
- Submitting institution
-
Queen's University of Belfast
- Unit of assessment
- 12 - Engineering
- Output identifier
- 161782451
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- AAAI-19: The Thirty-Third AAAI Conference on Artificial Intelligence: Proceedings
- First page
- 5361
- Volume
- -
- Issue
- -
- ISSN
- 2159-5399
- Open access status
- Compliant
- Month of publication
- -
- 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)
-
C - Electrical and Electronic
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Deep metric learning aims to capture the semantic similarity of data points. This paper identifies two critical limitations of the sample mining methods in deep metric learning, and provides general solutions for both of them. The work has been utilized by the company AnyVision in its various products, e.g., video surveillance and mobile device identification (contact: Guosheng Hu - guosheng.hu@anyvision.co).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -