Comparative Analysis of Active Contour Models for Brain Tumor segmentation from T1w MRI Images

Abstract

Early detection of brain tumors helps the doctors to treat the patients, also decrease the risk and improve the chance of survival. In the early days, brain tumor identification is being done manually, which need a lot of expertise and it's a time-consuming task. Magnetic resonance images (MRI) scans support getting brain tumors at an early stage, and it has been turned into a hot research topic in medical image processing. Hence, presented the comparative analysis of chan-vese (C-V) model and level set (LS) method intended to detect the tumor from human brain MRI images. This article's major objective is to compare the effectiveness of active contour models (ACM) intended to segment the brain tumor in terms of accuracy and computation time. To detect tumors, T1weighted abnormal images by the size of 256 × 256 were segmented after suppressing the noise by applying the pre-processing step. Both segmentation model's performance is assessed using the BRATS 2017 and Harvard university datasets. From all the responses, the C-V model is detecting the tumor effectively with a high speed as compared to the LS method. Finally, this work reveals that the C-V segmentation technique outer performs to detect the brain tumor with high accuracy and minimal processing time over the LS method.