An Effective System for Content MRI Brain Image Retrieval using Angular Radial Transform

Abstract

Nowadays, the growth of huge amount of medical image has become one of the most important clinical diagnosis components, Furthermore, there is an urgent need a system of Content Based Image Retrieval (CBIR) to obtain essential information such as type of image and extracting features of the image, such as color, shape and texture. This system is based on the image content that retrieves similar pathology involving magnetic resonance (MR) images of the medical database to assist the radiologist in the diagnosis of brain tumor; the shape recovery is the one of the top performers. Content-based image retrieval can also be used to locate brain tumors in medical images in large databases. In this paper, we propose a new method to build a new research system by the content of MRI images (CBIR) based on image characteristics, such as shape using the ART descriptor (angular radial transform) that is applied to reveal the characteristics of MR images. ART descriptor (angular radial transform) of shape is based Region adopted in MPEG-7 has the properties invariant to scale, rotation and robustness to noise, thanks to its properties we have used in this system. After the segmentation process, extracting the visual features of shape by calculating the coefficients of the ART and forming a second feature vector to be input to a support vector machine (SVM) for determining the presence tumor or not tumor followed by KNN (K-nearest neighbor) that retrieves the most similar images in the database. To provide faster image search. General TermsContent based image retrieval; Classification; KeywordsART descriptor (angular radial transform), Content based image retrieval (CBIR), magnetic resonance (MR) images, feature extraction, SVM, KNN.