A New Approach for Segmentation and Detection of Brain Tumor in 3D Brain MR Imaging

Abstract

Segmentation of brain MR Images plays prime role for measuring and visualizing the brain anatomical structures, analyzing brain tumors, and surgical planning. In the comparable research, outcomes were demonstrated the segmentation and detection of tumor in 2D Brain MR Images by amalgamation of diverse methods and techniques. Conversely, the precise outcomes were not been exhibited in the related researches works for the segmentation and detection of tumor in 3D Brain MR Images. As a result, this proposed work focused on development of an automatic integrated segmentation Frameworkfor detection of tumor in brain 3D MR Images which incorporate the most established improved EM (Expectation Maximization) method and Fuzzy C Means Clustering method. The proposed framework optimally merges the segmentation results of most established method and it exhibits the improvement in brain MR image segmentation. The most popular an anisotropic filter is employed to the improved EM (Expectation Maximization), Fuzzy C Means Clustering Method and Proposed Augmentation Method to improve the quality brain MR image and to produce better segmentation and detection of tumor. The performance results of proposed framework is evaluated on simulated brain Fluid-Attenuated Inversion Recovery MRI images and real brain dataset. The performance results of the proposed research work present superior results than the state-of-the-art methods and the proposed work is quantified with segmentation accuracy, sensitivity and specificity.