A 3-dimensional Brain Tumor Segmentation using Hybrid Intelligent Algorithms

Brain tumor is a cluster of tissue which is structured by continuing addition of abnormal or irregular cells and requires to be classified at its initial stage for treatment. Volumetric segmentation of brain tumor has an additional benefit in its analysis although, a tricky task in medical image processing. In that approach magnetic resonance imaging (MRI) has developed into a valuable medical diagnostic tool for the analysis of human brain for anomalies and discrepancies. In volumetric segmentation, complexity arises from the immense diversity in appearance and location and vast deviation in size and shape of tumors but acts as a crucial role in its quantitative investigation for treatment planning. The precision of segmentation scheme relies on its competence to differentiate unlike tissues on multiple modalities like T1-weighted, T2-weighted as well as T1-contrast enhanced images irrespective of having different contrasts for same tissue type. In this paper, we proposed a technique to carry out volumetric segmentation of brain tumor in three dimensions on several variations of brain MRI using a hybrid Confidence connected region growing algorithm and active contour model. The proposed technique is assessed and demonstrated that proposed technique achieves good result in similarity with ground truth and additional advanced schemes. The results presents 3D visualization of tumor volume that can support the practitioners in the identification of brain tumor for future treatment planning by providing information regarding tumor location and its features.