Image-Specific Classification With Local and Global Discriminations

doi:10.1109/TNNLS.2017.2748952

	Image-Specific Classification With Local and Global Discriminations
	Zhang, Chunjie1,2 ; Cheng, Jian2,3,4 ; Li, Changsheng 5; Tian, Qi 6
刊名	IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
	2018-09-01
卷号	29 期号:9 页码:4479-4486
关键词	Global information image-specific classification local information object categorization
ISSN号	2162-237X
DOI	10.1109/TNNLS.2017.2748952
通讯作者	Zhang, Chunjie(chunjie.zhang@ia.ac.cn)
英文摘要	Most image classification methods try to learn classifiers for each class using training images alone. Due to the interclass and intraclass variations, it would be more effective to take the testing images into consideration for classifier learning. In this brief, we propose a novel image-specific classification method by combing the local and global discriminations of training images. We adaptively train classifier for each testing image instead of generating classifiers for each class with training images alone. For each testing image, we first select its k nearest neighbors in the training set with the corresponding labels for local classifier training. This helps to model the distinctive characters of each testing image. Besides, we also use all the training images for global discrimination modeling. The local and global discriminations are combined for final classification. In this way, we could not only model the specific character of each testing image but also avoid the local optimum by jointly considering all the training images. To evaluate the usefulness of the proposed image-specific classification with local and global discrimination (ISC-LG) method, we conduct image classification experiments on several public image data sets. The superior performances over other baseline methods prove the effectiveness of the proposed ISC-LG method.
资助项目	National Natural Science Foundation of China[61303154] ; National Natural Science Foundation of China[61332016] ; Scientific Research Key Program of Beijing Municipal Commission of Education[KZ201610005012] ; ARO[W911NF-15-1-0290] ; National Science Foundation of China[61429201] ; Faculty Research Gift Awards by the NEC Laboratories of America ; Faculty Research Gift Awards by the NEC Laboratories of Blippar
WOS关键词	OBJECT CATEGORIZATION ; NEURAL-NETWORKS ; LOW-RANK ; CLASSIFIERS ; MODEL
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000443083700045
资助机构	National Natural Science Foundation of China ; Scientific Research Key Program of Beijing Municipal Commission of Education ; ARO ; National Science Foundation of China ; Faculty Research Gift Awards by the NEC Laboratories of America ; Faculty Research Gift Awards by the NEC Laboratories of Blippar
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/15318]
专题	自动化研究所_类脑智能研究中心
通讯作者	Zhang, Chunjie
作者单位	1.Chinese Acad Sci, Inst Automat, Ctr Brain Inspired Intelligence, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100049, Peoples R China 3.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, Beijing 2728, Peoples R China 4.Chinese Acad Sci, Ctr Excellence Brain Sci & Intelligence Technol, Beijing 2728, Peoples R China 5.Univ Elect Sci & Technol China, Sch Comp Sci & Engn, Chengdu 610051, Sichuan, Peoples R China 6.Univ Texas San Antonio, Dept Comp Sci, San Antonio, TX 78249 USA
推荐引用方式 GB/T 7714	Zhang, Chunjie,Cheng, Jian,Li, Changsheng,et al. Image-Specific Classification With Local and Global Discriminations[J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,2018,29(9):4479-4486.
APA	Zhang, Chunjie,Cheng, Jian,Li, Changsheng,&Tian, Qi.(2018).Image-Specific Classification With Local and Global Discriminations.IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS,29(9),4479-4486.
MLA	Zhang, Chunjie,et al."Image-Specific Classification With Local and Global Discriminations".IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 29.9(2018):4479-4486.