Content Based Medical Image Retrieval System using Multiple Classifier Framework

摘要

Content Based Image Retrieval (CBIR) is the set of techniques for searching of similar images from an image database using automatically extracted image features such as color, shape and texture. Feature extraction is the fundamental step in the CBIR systems. In the proposed multiple classifier method best discriminating features are combined in order to improve the retrieval performance. In the proposed system three shape features and three texture features are combined. The best matched features are found using the Bipartite weighted graph and combined using the decision combination operators. The retrieval experiments were conducted on the large medical image database and the results show that the proposed multiple classifier system yields good retrieval performance than the retrieval systems based on the individual features.