Segmentation and analysis of ventricles in Schizophrenic MR brain images using optimal region based energy minimization framework

Abstract

Schizophrenia (SZ) is a neurological disorder, which affects linguistic, memory, consciousness and executive functions of the brain. Magnetic resonance imaging (MRI) is used to capture structural abnormalities in human brain regions. In this work, segmentation of ventricle region from Schizophrenic MR brain images was carried out using optimized energy minimization framework. The images considered in this work are obtained from Centers of Biomedical Research Excellence (COBRE) database. Initially, the original images are subjected to simultaneous bias correction and segmentation using multiplicative intrinsic component optimization. The ventricles are extracted from other internal brain structures using this method. The obtained results are validated against the ground truth images. Results show that, multiplicative intrinsic component optimization method is able to segment ventricle from normal and SZ images. The correlation of ventricle area with ground truth is high (R = 0.99). It is noticed that SZ subjects have increased ventricle area compared to that of normal subjects. The high value of rand index (0.98) along with low value of global consistency error and variation of information shows the efficiency of the proposed method. The feature area extracted from the ventricle seems to be significant; hence it may be clinically supportive in the diagnosis of Schizophrenic subjects.