A RETROSPECTIVE FRAMEWORK FOR MEDICAL CBIR SYSTEM

Medical domain in the digital phase has increased its use of medical images from its defined limits to use in the diagnosis and the optimal treatment of the patients. The medical image retrieval approaches are reaching the extended versions. The present research work also has extended its active operation of feature vector by developed geometric approaches and calculated Edge Density in the concern of X-rays of five different human body parts. The intense analysis for the edge density has relate the relevant image for the query image with precision rate of 87.5% in contrast to 82.5 % achieved in the case of nearest neighbor and 85% in the Five nearest neighbor. It also covers the comparison with the techniques involving either single feature or multiple feature in the related researches for the retrieval of medical images to its geometric feature vectors used in the more appropriate or most relevant retrieval image for the system