A two-stage hybrid probabilistic topic model for refining image annotation

doi:10.1007/s13042-019-00983-w

	A two-stage hybrid probabilistic topic model for refining image annotation
	Tian, Dongping 1; Shi, Zhongzhi 2
	2020-02-01
发表期刊	INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS
ISSN	1868-8071
卷号	11 期号:2 页码:417-431
摘要	Refining image annotation has become one of the core research topics in computer vision and pattern recognition due to its great potentials in image retrieval. However, it is still in its infancy and is not sophisticated enough to extract perfect semantic concepts just according to the image low-level features. In this paper, we propose a two-stage hybrid probabilistic topic model to improve the quality of automatic image annotation. To start with, a probabilistic latent semantic analysis model with asymmetric modalities is learned to estimate the posterior probabilities of each annotation keyword, during which the image-to-word relation can be well established. Next, a label similarity graph is constructed by a weighted linear combination of label similarity and visual similarity of images associated with the corresponding labels. By this way, the information from image low-level visual features and high-level semantic concepts can be seamlessly integrated by fully taking into account the word-to-word and image-to-image relations. Finally, the rank-two relaxation heuristics is exploited to further mine the correlation of the candidate annotations so as to capture the refining results, which plays a critical role in semantic based image retrieval. Extensive experiments show that the proposed model achieves not only superior annotation accuracy but also better retrieval performance.
关键词	Refining image annotation Semantic gap Expectation-maximization PLSA Max-bisection Image retrieval
DOI	10.1007/s13042-019-00983-w
收录类别	SCI
语种	英语
资助项目	National Program on Key Basic Research Project (973 Program)[2013CB329502] ; National Natural Science Foundation of China[61035003] ; National Natural Science Foundation of China[61202212] ; Tianchenghuizhi Fund for Innovation and Promotion of Education[2018A03036] ; Key R&D Program of the Shaanxi Province of China[2018GY-037]
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000512019400011
出版者	SPRINGER HEIDELBERG
引用统计	被引频次：12[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/14776
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Tian, Dongping
作者单位	1.Baoji Univ Arts & Sci, Inst Comp Software, Baoji 721007, Shaanxi, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Tian, Dongping,Shi, Zhongzhi. A two-stage hybrid probabilistic topic model for refining image annotation[J]. INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS,2020,11(2):417-431.
APA	Tian, Dongping,&Shi, Zhongzhi.(2020).A two-stage hybrid probabilistic topic model for refining image annotation.INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS,11(2),417-431.
MLA	Tian, Dongping,et al."A two-stage hybrid probabilistic topic model for refining image annotation".INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 11.2(2020):417-431.