Target Detection from Brain MRI and Its Classification

Early detection of brain tumor is a very exigent task for radiologists. Delineating the boundaries of brain tumors from magnetic resonance imaging is an important task for brain cancer research. Throughout the most recent decade, various strategies have just been proposed for automated brain tumor detection. In this paper, we have introduced a novel strategy to characterize a given MR cerebrum image as normal or having unusual characteristics. The proposed strategy incorporates automated segmentation using K-means clustering, extraction of textural features using discrete wavelet transformation trailed by applying principle component analysis (PCA) to reduce the dimension of elements of highlights. The diminished highlights were submitted to a kernel support vector machine (KSVM) which was tested using three different kernels. It could be applied to the field of MR brain image characterization and can assist the specialists to analyze the degree of abnormality and severity of the brain tumor. Experimental findings indicate that the proposed approach offers an efficient and promising tool for highlighting and classifying brain tumors from MR images with an average accuracy of 79%.