Hippocampus Segmentation Using Fuzzy C Means-Based Level Set Local Ternary Pattern with Enhanced Edge Indicator

The hippocampus segmentation plays a crucial role in analyzing brain activities, which is a widely used biomarker for Alzheimer’s disease, epilepsy, and schizophrenia diagnosis. The automatic segmentation of hippocampus is a complex task, because of low signal contrast, small structural size, and insufficient image resolution. The automatic hippocampus segmentation utilizing magnetic resonance imaging (MRI) is effective in clinical diagnosis and neuro-science research. In this manuscript, a new automatic segmentation model in MRI is implemented for hippocampus segmentation. Initially, the MRI brain images are collected from Neuroimaging Tools and Resources Collaboratory (NITRC) and the Open Access Series of Imaging Studies (OASIS) databases. Further, color normalization technique is employed for improving the visual ability, and reduce the impulse noise and machinery noise (mechanical and electrical noises) in the image. Finally, the Fuzzy C Means-based Level Set Local Ternary Pattern with Enhanced Edge Indicator (FCM-LSLTPEEI) is proposed for hippocampus segmentation. The proposed model combines FCM and EEI functions in the LSLTP model, where the important phase is to adapt the EEI function effectively with the LSLTP. The experimental outcomes revealed that the proposed FCM-LSLTPEEI model obtained 98.90% and 99.01% of accuracy on the OASIS and NITRC databases, which are superior compared to the traditional models.