Optimal Kernels for Neuromorphic Soft Sets in Brain Image Segmentation using Pareto-Based Clustering Algorithm

Research on biomedical image processing proliferates specifically in brain image segmentation. Cerebral disorders often present with the formation of tumors and necrotic tissues which require accurate tissue extraction and classification from brain-MRI. Therefore, mathematical approaches like the fuzzy sets, rough sets or soft set can then measure the amount of uncertainty in the brain images. This in turn improves segmentations accuracy. Asymmetric classification of brain tumour continues to emerge as a problem due to the improve quality, reduction of noise and image segmentation. The work being presented in this paper is a new approach to segmentation of brain images using neuromorphic soft sets. This approach tries to model uncertainty in white matter, grey matter as well as the cerebrospinal fluid. This technique helps find new brain malignancies in the magnetic resonance imaging. In this context, we promote Pareto Optimization for the identification of the best values of kernel and functions and neuromorphic soft set approximations as well. These approximations are comparatively easier than traditional algorithms because the latter involve no interaction from the user side. They measure weight, average, median, and other variables. The comprehensive experiments confirmed that the proposed method is faster in terms of processing and testing jacquard parameters and more accurate in segmenting organs compared to the existing works.