首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于视觉注意机制的小目标预检测特性
引用本文:王岳环,路自谦,张天序.基于视觉注意机制的小目标预检测特性[J].华中科技大学学报(自然科学版),2006,34(8):10-11,25.
作者姓名:王岳环  路自谦  张天序
作者单位:华中科技大学,图像识别与人工智能研究所,湖北,武汉,430074
基金项目:国家自然科学基金;国防科技预研项目
摘    要:针对一种使用了离散非线性采样的视觉注意机制模型,对其应用在小目标预检测上的检测概率进行了研究和分析.在一定的假设条件下,从理论上推导出了模型中影响小目标预检测概率的各个因素,包括目标大小、目标所在采样节点的大小、输入图像信噪比和目标均匀性参数等.由于采用了非线性的采样模型,注视点的转移将带来不同图像区域空间分辨率的变化,从而不能获得总的检测概率与上述各参数的有意义的关系,通过实验分析了这些因素在小目标预检测中对检测概率的具体影响.

关 键 词:小目标预检测  离散非线性采样  性能分析  注意机制
文章编号:1671-4512(2006)08-0010-02
收稿时间:2006-03-06
修稿时间:2006年3月6日

Performance analysis of small target pre-detection based on visual attention mechanism
Wang Yuehuan,Lu Ziqian,Zhang Tianxu.Performance analysis of small target pre-detection based on visual attention mechanism[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2006,34(8):10-11,25.
Authors:Wang Yuehuan  Lu Ziqian  Zhang Tianxu
Abstract:A visual attention mechanism model for discrete non-linear sampling is studied to analyze its detection probability during pre-detecting small targets. Based on some reasonable hypothesizes, the factors influencing the detection probability were deduced, including the target size, the size of the sampling node which contains the target, the signal-noise ratio of the input image and the uniformity of the target. As a result of the non-linear sampling model and change of the space resolution in different regions of the image brought by the transition of the gaze point, a specific relationship between the overall detection probability and the various factors could not be deduced, some experiments were performed to find the detailed effects of these factors on the de.tection probability.
Keywords:small target detection  discrete non-linear sampling  performance analysis  attention mechanism
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号