摘要
基于关键词的网络图像检索得到的结果往往是海量且无序的,给用户造成不便。本文构建了一个网络图像检索结果重排序系统并进行了仿真。提出了一种利用图像SIFT局部特征构建图像间视觉Graph结构对图像进行聚类的重排序算法,并引入检索关键词的网络翻译以提高检索效率。仿真结果证明,系统的重排序结果满足网络用户的检索期望,验证了算法的有效性。
The results of web image retrieval based on key-word query tend to be huge in quantity and disordered, which causes much inconvenience to users. In this paper, a re-ranking system of web image retrieval results is constructed and simulation is made. Also a re-ranking algorithm is proposed, which use SIFT local features of the images to construct visual graph among images so as to cluster the images, in the same time Web Explanation of keywords is introduced to enhance the efficiency of retrieval. Experiments show that re-ranking results are well meeting the expectation of web users, and the efficiency of the algorithm is verified.
出处
《电子技术（上海）》
2009年第11期47-49,46,共4页
Electronic Technology
