10.22376/ijpbs.2019.10.1.p1-12 Volume 5 Issue 4 2014 (October - December) Medical image retrieval based on user query is a challenging task. A number of methods exist which use text or content of related image for retrieval. In case of content based retrieval, the various features of an image such as color, shape and text from an user perspective are analyzed and those features are searched through the process of segmentation. This type of search may not be suitable for medical images because the accurate difference between images can not be identified with the normal human eye. In this paper, we propose a method, a combination of the Statistical Region Merging (SRM), Discrete Cosine Transform (DCT) and RGB projection to deal with the similarity between images, the related images of query image are retrieved with similarity percentage and displayed from view. The experimental analysis of this paper shows the accuracy of the proposed method is good compared to other existing features extraction methods. GOBIGA JEGATHEESAN , ANUSUYA S. VENKATESAN AND S. SATHISH CBIR, DCT, RGB projection, SRM and Image Retrieval. 237-242