Optimizing top precision performance measure of content-based image retrieval by learning similarity function

Handle URI:
http://hdl.handle.net/10754/623313
Title:
Optimizing top precision performance measure of content-based image retrieval by learning similarity function
Authors:
Liang, Ru-Ze; Shi, Lihui; Wang, Haoxiang; Meng, Jiandong; Wang, Jim Jing-Yan; Sun, Qingquan; Gu, Yi
Abstract:
In this paper we study the problem of content-based image retrieval. In this problem, the most popular performance measure is the top precision measure, and the most important component of a retrieval system is the similarity function used to compare a query image against a database image. However, up to now, there is no existing similarity learning method proposed to optimize the top precision measure. To fill this gap, in this paper, we propose a novel similarity learning method to maximize the top precision measure. We model this problem as a minimization problem with an objective function as the combination of the losses of the relevant images ranked behind the top-ranked irrelevant image, and the squared Frobenius norm of the similarity function parameter. This minimization problem is solved as a quadratic programming problem. The experiments over two benchmark data sets show the advantages of the proposed method over other similarity learning methods when the top precision is used as the performance measure.
KAUST Department:
Physical Sciences and Engineering (PSE) Division; Materials Science and Engineering Program
Citation:
Ru-Ze Liang, Shi L, Wang H, Jiandong Meng, Wang JJ-Y, et al. (2016) Optimizing top precision performance measure of content-based image retrieval by learning similarity function. 2016 23rd International Conference on Pattern Recognition (ICPR). Available: http://dx.doi.org/10.1109/ICPR.2016.7900086.
Publisher:
IEEE
Journal:
2016 23rd International Conference on Pattern Recognition (ICPR)
Issue Date:
24-Apr-2017
DOI:
10.1109/ICPR.2016.7900086
Type:
Conference Paper
Sponsors:
The study is supported by a grant from Provincial Key Laboratory for Computer Information Processing Technology, Soochow University, China
Additional Links:
http://ieeexplore.ieee.org/document/7900086/
Appears in Collections:
Conference Papers; Physical Sciences and Engineering (PSE) Division; Materials Science and Engineering Program

Full metadata record

DC FieldValue Language
dc.contributor.authorLiang, Ru-Zeen
dc.contributor.authorShi, Lihuien
dc.contributor.authorWang, Haoxiangen
dc.contributor.authorMeng, Jiandongen
dc.contributor.authorWang, Jim Jing-Yanen
dc.contributor.authorSun, Qingquanen
dc.contributor.authorGu, Yien
dc.date.accessioned2017-05-02T13:22:29Z-
dc.date.available2017-05-02T13:22:29Z-
dc.date.issued2017-04-24en
dc.identifier.citationRu-Ze Liang, Shi L, Wang H, Jiandong Meng, Wang JJ-Y, et al. (2016) Optimizing top precision performance measure of content-based image retrieval by learning similarity function. 2016 23rd International Conference on Pattern Recognition (ICPR). Available: http://dx.doi.org/10.1109/ICPR.2016.7900086.en
dc.identifier.doi10.1109/ICPR.2016.7900086en
dc.identifier.urihttp://hdl.handle.net/10754/623313-
dc.description.abstractIn this paper we study the problem of content-based image retrieval. In this problem, the most popular performance measure is the top precision measure, and the most important component of a retrieval system is the similarity function used to compare a query image against a database image. However, up to now, there is no existing similarity learning method proposed to optimize the top precision measure. To fill this gap, in this paper, we propose a novel similarity learning method to maximize the top precision measure. We model this problem as a minimization problem with an objective function as the combination of the losses of the relevant images ranked behind the top-ranked irrelevant image, and the squared Frobenius norm of the similarity function parameter. This minimization problem is solved as a quadratic programming problem. The experiments over two benchmark data sets show the advantages of the proposed method over other similarity learning methods when the top precision is used as the performance measure.en
dc.description.sponsorshipThe study is supported by a grant from Provincial Key Laboratory for Computer Information Processing Technology, Soochow University, Chinaen
dc.publisherIEEEen
dc.relation.urlhttp://ieeexplore.ieee.org/document/7900086/en
dc.rightsArchived with thanks to 2016 23rd International Conference on Pattern Recognition (ICPR)en
dc.subjectBenchmark testingen
dc.subjectComputersen
dc.subjectImage retrievalen
dc.subjectLearning systemsen
dc.subjectOptimizationen
dc.subjectTrainingen
dc.titleOptimizing top precision performance measure of content-based image retrieval by learning similarity functionen
dc.typeConference Paperen
dc.contributor.departmentPhysical Sciences and Engineering (PSE) Divisionen
dc.contributor.departmentMaterials Science and Engineering Programen
dc.identifier.journal2016 23rd International Conference on Pattern Recognition (ICPR)en
dc.eprint.versionPost-printen
dc.contributor.institutionCenterfield Corporation, Los Angeles, CA 90245, USAen
dc.contributor.institutionDepartment of Electrical and Computer Engineering, Cornell University, Ithaca, NY 14850, USAen
dc.contributor.institutionShandong Medical College, Linyi, 276000, Chinaen
dc.contributor.institutionNew York University Abu Dhabi, United Arab Emiratesen
dc.contributor.institutionSchool of Computer Science and Computer Engineering, California State University, San Bernardino, 92407, USAen
dc.contributor.institutionTravelers Canada, Toronto, ON, Canadaen
kaust.authorLiang, Ru-Zeen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.