Automatic localization and level set based energy minimization for MRI brain tumor

Abstract

Automatic segmentation of tumor abnormality is a very difficult task for the radiologist. In this research, we proposed a located brain tumor with automatic seed point localization and no need to initially select the location of the region which is to be infected. Estimation of the abnormalities for initial bounding box after this, we proposed the segmentation of tumor called automatic level set minimization function with a new technique that is localization based energy minimization of MRI brain tumor. The performance of localization is evaluated using based on the level of detection and radiologist analytical results. Total 100 FLAIR, T1, and T2-weighted MRI brain tumor images (Astrocytoma (22), Ganglioglioma (6), Glioblastoma (23), Epidermoide (3), Mixed Glioma (5) and Meningnet (41)) (5type of tumors) were used for the experiment. Experimental results show that the method has successfully localized the brain tumors with 97% accuracy.