Abstract: For similar medical case searching in computer-aided diagnosis,a multi-modality information retrieval algorithm based on multi-graph semi-supervised learning model is proposed.This paper first regards medical images and texts as compensatory multi-modality information.The graph models for both kinds of information are constructed respectively and a multi-graph fusion framework is defined.Then,a manifold ranking procedure is generated based on this fusion graph to obtain the final medical case retrieval results.Experimental results on mammograph dataset and lung CT dataset demonstrate that the proposed algorithm can effectively utilize the image and textual information for medical case retrieval,and achieves a more effective and efficient retrieval performance.