DCT-Inspired Feature Transform for Image Retrieval and Reconstruction
- Submitting institution
-
King's College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 128739039
- Type
- D - Journal article
- DOI
-
10.1109/TIP.2016.2590323
- Title of journal
- IEEE TRANSACTIONS ON IMAGE PROCESSING
- Article number
- -
- First page
- 4406
- Volume
- 25
- Issue
- 9
- ISSN
- 1057-7149
- Open access status
- Technical exception
- Month of publication
- July
- 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
-
3
- Research group(s)
-
-
- Citation count
- 16
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper revisits visual feature extraction and description, which is the cornerstone of many computer vision applications. It formulates a novel way to determine the dominant orientation for feature extraction and a new DCT inspired feature transform (DIFT). Extensive experiments on both image retrieval and reconstruction tasks with several benchmarks are provided to show that the feature redundancy is reduced by 40% and the efficiency improves threefold compared to the state-of-the-art (e.g. SIFT). The way the DCT component is formulated and interpreted has been adopted in subsequent works, e.g. [Wang, Xu, You, TIP2017], [Wang, Xu, Xu, AAAI2017].
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -