Abstract

: Brain tumor (BT) segmentation is a significant process in clinical imaging. Earlier identification of BT acts as a crucial process to improve the treatment prospects and enhances the endurance rate of the patients. Manual segmentation of the BTs for cancer detection from enormous quantity of MRI images produced in a daily medical schedule is a troublesome and tedious process. Therefore, an automated BT segmentation tool is required. In this view, this paper presents a novel Likelihood based Probability Density Function with Relevance Vector Machine (LPDF-RVM) and Growing Convolution Neural Network (GCNN) for automatic BT segmentation. Initially, MR brain images are taken as input. In addition, the image normalization process and wiener filter based noise removal process takes place along with the skull stripping process.  After that, Gray-Level Co-Occurrence Matrix (GLCM) and Histogram of Oriented Gradients (HOG) features extraction is performed. Based on the features, LPDF-RVM is applied as a segmentation technique followed by GCNN. A detailed simulation analysis was carried out to highlight the better performance of the presented model.