视觉选择性注意的模型化计算及其应用前景  

The Modeling of Visual Selective Attention and Its Applications

作　　者：刘琼[1] 秦世引[1] 李志成[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100191

出　　处：《科技导报》2010年第1期107-115,共9页Science & Technology Review

摘　　要：视觉选择性注意(Visual Selective Attention,VSA)的模型化计算是当今认知信息处理领域的热点前沿研究方向,在计算机智能视觉感知领域有着广阔的应用前景。对近年来视觉选择性注意研究领域的成果进行了回顾,着重对中央区自下而上兴趣圈的产生、视觉任务和物体知识的表达、自上而下与自下而上的双向信息流的汇合以及显性注意分支中凝视点的转移控制等问题中具有较高影响力的模型化计算方法的效能和特点进行了分析与探讨,对相关争论进行了归结和评述。进而通过VSA模型化计算在工程应用领域特别是视频压缩、目标检测、广告设计等方面的成功应用范例展示了其工程应用价值,最后对其应用前景进行了展望。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.

关 键 词：认知信息处理 主动视觉 视觉选择性注意 模型化计算 

分 类 号：B842[哲学宗教—基础心理学]