Region Based Multiple Features for an Effective Content Based Access Medical Image Retrieval an Integrated with Relevance Feedback Approach

Abstract

An efficient Content-based medical image retrieval (CBMIR) system is imperative to browse the entire database to locate required medical image. This paper proposes an effective scheme includes the detection of the boundary of the image followed by exploring the content of the interior boundary region with the help of multiple features. The proposed technique integrates the Texture, Shape features and the relevance feedback mechanism. Differentiate of Gabor Filter used for Texture feature extraction and Moments extract the Region based shape features. The Euclidean distance is used for similarity measure and then these distances are sorted out and ranked. The Recall rate of the medical retrieval system has been enhanced by adapting Relevance Feedback mechanism. The efficiency of the proposed method has been evaluated by using a huge data base by employing multiple features and integrating with Relevance feedback approach. Correspondingly, the Recall Rate has been enormously enhanced and Error Rate has been reduced as compared to the existing classical retrieval methods.