Abstract: To promote the medical image retrieval performances based on content, this paper integrates image retrieval based on the fuzzy regional feature into the relevance feedback algorithm based on support vector machine, and proposes a new joint relevance feedback algorithm. This algorithm uses fuzzy regional feature based on the algorithm for regional feature, the global wavelet energy features use both of the relevant images and the irrelevant images based on relevance feedback algorithm of SVM. Experiment shows that the hybrid algorithm performs better.

Key words: relevance feedback, image retrieval, SVM