REGRESSIVE NONLINEAR TEAGER FILTER BASED MAP ESTIMATED RELEVANCE VECTOR SEGMENTATION FOR BRAIN MRI IMAGE PROCESSING

Abstract

Magnetic resonance images (MRI) is significant in medical diagnosis as it provides detailed information related to anatomical structures as well as abnormal tissues of the body for treatment planning. The current medical imaging research is still a very difficult task to diagnosis the disease perfectly. Since the developed imaging system has more error for exact analysis. In order to overcome such issues, Regressive Nonlinear Teager Filter based MAP Estimated Relevance Vector Image Segmentation (RNTF-MAPRVIS) Method is developed for processing the brain MRI images with higher accuracy and minimum time. The numbers of brain MRI images are collected from the database. The RNTF-MAPRVIS method performs two major processes with medical images, namely preprocessing and segmentation. Initially, the Regressive Nonlinear Teager Filter process is used to remove the noisy pixels from the image. The designed filter analyzes the relationship of an image pixel to obtain super-resolution brain MR image through the warping and interpolation. After preprocessing, MAP estimated Relevance Vector Machine based image segmentation process is carried out to segment the input preprocessed image for finding as normal or abnormal. In RNTF-RVIS Method, Relevance Vector Machine constructs the hyperplane uses Maximum a Posteriori that segments the images based on the similarity between the extracted features and testing features. After performing the segmentation, the input image is said as normal or abnormal. Experimental evaluation is carried out on factors such as PSNR, segmentation accuracy, segmentation time and false positive rate with respect to a number of MRI images. The observed results prove that the presented RNTF-MAPRVIS method improves the segmentation accuracy, PSNR and minimize time as well as the false positive rate than the state-of-the-art methods.