基于感知的大范围人造对象图像结构特征抽取

基于内容的图像检索(CBIR)系统传统上仅仅使用颜色特征和纹理特征进行图像的底层分析,不足以满足检索性能的要求.要提高基于内容图像检索的鲁棒性,需要检索系统能分析更多的图像底层特征.对于包含大范围人造对象(比如:建筑物、高塔、桥梁和其他结构对象)的图像,提出应用感知分组(perceptual grouping)规则抽取图像的结构特征,并结合使用该结构特征,对特定的图像库进行分类和客观评价.

Structure feature extraction of images containing large manmade objects based on perceptual grouping

Content-based image retrieval(CBIR) systems have traditionally used color feature and texture feature for low-level image analyses,which cannot meet wholly with people's performance need.To improve the need for robust of CBIR,it is very necessary to analyze more low-level image features in the image retrieval system.The paper applies perceptual grouping rules to extract structure feature of images containing large manmade objects such as buildings,towers,bridges,and other architectural objects.At last,structural feature is employed to classify special image database and valued a little.