Medical Image Feature Classification from Rule Mining and Retrieval via Weighted Swarm Neural Network

Content Based Medical Image Retrieval (CBMIR) is a system to browse, search, and retrieve medical images similar to the query image. Since most of the medical images are gray scale images, it is important to concentrate on the accuracy criteria to support the diagnosis task of medical practitioners. In this paper, we propose medical image retrieval with data mining and Neural Network. The data-mining algorithm detects the contents of the image in query and database images. The problem of negative retrieval in medical images is due to the lack of cognition. The lack of cognitions in retrieval performs more negative retrieval images. In negative image retrieval, for lung query image the retrieval give the equivalent images with lungs as well as tissue images. To avoid negative image retrieval, we apply Context based Association Rules Mining to eliminate the irrelevant images during retrieval process. However, the lack of cognitions solves with CBIR and the priority and pattern recognition for equivalent images obtain through weighted swarm Neural Network. As a result, the retrieval system discussed here takes its advantage over existing system through its high accuracy and minimal retrieval time.