A New Method of Content Based Medical Image Retrieval and Its Applications to CT Imaging Sign Retrieval.
Ling Ma,Xiabi Liu,Yan Gao,Yanfeng Zhao,Xinming Zhao,Chunwu Zhou
DOI: https://doi.org/10.1016/j.jbi.2017.01.002
IF: 8
2017-01-01
Journal of Biomedical Informatics
Abstract:This paper proposes a new method of content based medical image retrieval through considering fused, context-sensitive similarity. Firstly, we fuse the semantic and visual similarities between the query image and each image in the database as their pairwise similarities. Then, we construct a weighted graph whose nodes represent the images and edges measure their pairwise similarities. By using the shortest path algorithm over the weighted graph, we obtain a new similarity measure, context-sensitive similarity measure, between the query image and each database image to complete the retrieval process. Actually, we use the fused pairwise similarity to narrow down the semantic gap for obtaining a more accurate pairwise similarity measure, and spread it on the intrinsic data manifold to achieve the context-sensitive similarity for a better retrieval performance. The proposed method has been evaluated on the retrieval of the Common CT Imaging Signs of Lung Diseases (CISLs) and achieved not only better retrieval results but also the satisfactory computation efficiency.