基于生物视觉感知的场景特征提取方法  

Scene Feature Extraction Method Based on Biological Visual Perception

作　　者：陈硕[1] 吴成东[1] 陈东岳[1] 楚好[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110819

出　　处：《东北大学学报（自然科学版）》2011年第9期1217-1220,1228,共5页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(60874103;61005032)

摘　　要：针对复杂背景环境下图像场景理解及分类中遇到的诸多难题,提出了一种模拟生物视觉感知的新的图像场景特征提取方法.该方法采用方向、颜色和亮度三种感知特征来构造视觉显著图,并对基本单元进行最大化处理来得到更加精确的方向信息.为了提高特征提取的速度,设计了区域分割算法,然后构造分割区域描述子,最后通过SVM对特征向量集合进行无监督分类以实现显著区域匹配.与传统方法相比,该方法提取的特征对于一定程度的仿射和光照变化具有良好的不变性,同时具有较好的分类效果.实验结果表明上述方法能够较好地应用于场景分类及检索,并具有较强的时效性.According to the problems encountered in image understanding and classifying in the complex environment,a novel scene feature extraction method was proposed.This method adopts perceptive features of orientation,color and intensity to construct saliency map and the maximum operation is used to get the more accurate orientation information.A region segmentation algorithm is designed to accelerate feature extraction.Then,region segmentation descriptor is constructed.Finally,image matching is implemented through the support vector machine（SVM）.Compared with traditional methods,it has a good invariance in image scaling,translation and robustness across a substantial range of affine distortion.Meanwhile,it has good classification results.The experimental results demonstrate that the method is well applied to scene classification and retrieval with better timeliness.

关 键 词：场景分类 生物视觉感知 显著图 区域分割 区域描述子 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]