利用二部图匹配进行图像相似性度量

基于内容图像检索是多媒体信息检索领域研究的热点，而现有的算法和系统离成熟的应用还相距甚远，其检索效率和准确性都相当低。提高基于内容图像检索性能的关键在于实现对图像的对象级访问，但是已有的很多的基于区域的图像检索算法和系统都没有考虑多区域的匹配问题，因而不具有一般性、实用性。文中提出一种基于二部图最大权匹配的图像相似性度量算法，该算法建立在图像分割的基础上，由于它能有效地解决多区域图像相似性度量问题，并能有效地避免由于分割不准确带来的影响，因此能极大地提高检索的相关性和准确性。

Image Similarity Measure Using Max Weighted Bipartite Matching

Recently content-based retrieval of images and video has become a hot research area in the world, however, existing algorithms and systems can't meet users' requirements because of its low efficiency and low precision. The key to effectively improve the CBIR performance lies in the ability to access the image at the level of objects. Although many region-based CBIR algorithms and systems are proposed, multi-region in segmented images is still not taken into consideration, so they are not comprehensive and reasonable. In this paper, we present a framework of region-based image retrieval system using max weighted bipartite matching based on image segmentation. Because of taking spatial information into consideration and combining the information from all image regions, new algorithm can work effectively and efficiently.