Probabilistic Multi-Task Learning for Visual Saliency Estimation in Video

doi:10.1007/s11263-010-0354-6

	Probabilistic Multi-Task Learning for Visual Saliency Estimation in Video
	Li, Jia 2,3; Tian, Yonghong 1; Huang, Tiejun 1; Gao, Wen 1
	2010-11-01
发表期刊	INTERNATIONAL JOURNAL OF COMPUTER VISION
ISSN	0920-5691
卷号	90 期号:2 页码:150-165
摘要	In this paper, we present a probabilistic multi-task learning approach for visual saliency estimation in video. In our approach, the problem of visual saliency estimation is modeled by simultaneously considering the stimulus-driven and task-related factors in a probabilistic framework. In this framework, a stimulus-driven component simulates the low-level processes in human vision system using multi-scale wavelet decomposition and unbiased feature competition; while a task-related component simulates the high-level processes to bias the competition of the input features. Different from existing approaches, we propose a multi-task learning algorithm to learn the task-related "stimulus-saliency" mapping functions for each scene. The algorithm also learns various fusion strategies, which are used to integrate the stimulus-driven and task-related components to obtain the visual saliency. Extensive experiments were carried out on two public eye-fixation datasets and one regional saliency dataset. Experimental results show that our approach outperforms eight state-of-the-art approaches remarkably.
关键词	Visual saliency Probabilistic framework Visual search tasks Multi-task learning
DOI	10.1007/s11263-010-0354-6
收录类别	SCI
语种	英语
资助项目	Chinese National Natural Science Foundation[60973055] ; Chinese National Natural Science Foundation[90820003] ; National Basic Research Program of China[2009CB320906] ; Fok Ying Dong Education Foundation[122008]
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:000281087900002
出版者	SPRINGER
引用统计	被引频次：104[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/2XEOYT63/12507
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, Yonghong
作者单位	1.Peking Univ, Natl Engn Lab Video Technol, Beijing 100871, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Key Lab Intelligent Informat Proc, Beijing 100190, Peoples R China 3.Chinese Acad Sci, Grad Univ, Beijing 100049, Peoples R China
推荐引用方式 GB/T 7714	Li, Jia,Tian, Yonghong,Huang, Tiejun,et al. Probabilistic Multi-Task Learning for Visual Saliency Estimation in Video[J]. INTERNATIONAL JOURNAL OF COMPUTER VISION,2010,90(2):150-165.
APA	Li, Jia,Tian, Yonghong,Huang, Tiejun,&Gao, Wen.(2010).Probabilistic Multi-Task Learning for Visual Saliency Estimation in Video.INTERNATIONAL JOURNAL OF COMPUTER VISION,90(2),150-165.
MLA	Li, Jia,et al."Probabilistic Multi-Task Learning for Visual Saliency Estimation in Video".INTERNATIONAL JOURNAL OF COMPUTER VISION 90.2(2010):150-165.