一种基于视觉单词的图像检索方法

摘要

基于内容的图像检索技术最主要的问题是图像的低层特征和高层语义之间存在着“语义鸿沟”.受文本内容分析的启发,有研究学者借鉴传统词典中用文本单词组合解释术语的思路,将图像视为视觉单词的组合,利用一系列视觉单词的组合采描述图像的语义内容.为此,利用SIFT进行图像的视觉单词特征提取,然后构建视觉单词库,最后实现了一个基于视觉单词的图像检索系统.实验结果表明,该方法在一定程度上提高了图像检索的查准率.%The main problem of content-based image retrieval ( CBIR) is the semantic gap between the low level features and high-level semantics of image. Desired by text content analysis which it explains terminology by combining several text words, some researchers regard image as several visual words to describe the semantic content of images. To this end, SIFT (scale-invariant feature transform) algorithm is used to extract image features and then a dictionary of visual word is constructed. Finally, an image retrieval system based on visual words is realized. Experimental results show that the proposed approach can improves the precision rate of image retrieval to some extent.