An Effective Hybrid Approach For Extracting Abnormal Regions In Mr Brain Images With Deep Networks

Abstract

The functional framework of the anatomical work is acquired from segmented brain region; this is essential in neuro science research works. Abnormal region extraction through segmentation is a primary process in assessing the extent of disease. Automatic Brain Tumor segmentation using image processing techniques are employed in extracting these abnormal regions. This paper proposed a hybrid approach of brain segmentation using customized convolutional neural network and extraction of abnormal region with co-clustering algorithm. Metrical analysis prove that this approach is far superior than earlier methods.