A New Method for Content-Based Image Retrieval via Subsets of Key Contours Fragments
In this paper, a novel method is proposed for solving an open problem of shape matching in content-based image retrieval. In order to prepare for matching two images, our novel method uses an improved segment method to get more accurate contours. And then we find a new way to extract the key contour fragments, so that a shape can be presented as many key contour fragments in this way. It is an effective way to do analysis in image contour. In addition, a similar measure method is proposed to get the correlation rate of two images via subsets of key contour fragments. Finally, experiment results show that our new method can get more robustness than existing methods and achieve a superior matching effect in the standard shape databases MPEG-7.