Hybrid segmentation technique for meningioma tumour detection in MRI brain images

To ameliorate the measure of segmentation methods, a contrastive strategy of the hybrid segmentation is suggested in this report. A medical image is complex and difficult to diagnose for the disease identification, surgical preparation, and treatment. Magnetic Resonance Imaging (MRI) of the brain is useful for evaluating problems in the presence of an abnormality in the brain. Two different segmentation approaches are presented in this formulation which uses medical images to produce a hybrid approach. The single-seeded region based segmentation is the disunion of an image into homogenous parts of linked pixels. It develops the region, according to an iterated process and examines the neighboring pixels, whether they should be summarized within the region or not. Another method employed in this work is the thresholding image segmentation using the Differential Evolution (DE) based on entropy parameters. DE is a population-based, effectual and direct search method. It was designated because of its propensity to offer fast convergence rate and capacity of traveling straight away with real numbers (gray scale levels). In this study, the MRI image goes look over a proper pre-processing, like skull stripping and enhancement. After the aforementioned methods are enforced along these images and combine together, the resultant images are analyzed using Structure Similarity Index Measure (SSIM) to obtain a better value, when compared with the ground truth image