The Modeling of Visual Selective Attention and Its Applications

Visual Selective Attention(VSA) modeling is an advanced topic in cognitive information perception. It bridges neurophysiology,psychology and computer vision,and it aimed at developing a human-like visual system for computer. It shows a wide and promising application in intelligent computer vision from the late of last century. Firstly,recent VSA research topics in VSA modeling area are introduced in this paper,while the important and key problems are analyzed and discussed in details. In these topics,origin of bottom-up interesting map,representation of visual task and object,combination of bottom-up and top-down information,and fixation shifting in overt attention are taken as the highly influential aspects and thus described elaborately. The controversial issues in these topics are discussed as well. Secondly,in the view of application,the significance of VSA modeling are demonstrated by introducing and analyzing its successful application instances in video compression,target detection,and advertisement design,. The inevitable problems and challenges of VSA modeling are summarized and its possible future applications from VSA modeling are prospect in the end.