Optimal Kernel based Neutrosophic Soft Sets Clustering for Image Segmentation based on Pareto Optimal Algorithm
B. Prasanthi1, N. Nagamalleswararao2

1B. Prasanthi, Research Scholar, ANU, Assistant Professor, Department of IT, R.V.R & J.C College of Engineering, Guntur (Andhra Pradesh), India.
2Dr. N. Nagamalleswararao, Professor, Department of IT, R.V.R & J.C College of Engineering, Guntur (Andhra Pradesh), India.
Manuscript received on 26 March 2019 | Revised Manuscript received on 07 April 2019 | Manuscript Published on 18 April 2019 | PP: 810-818 | Volume-7 Issue-6S March 2019 | Retrieval Number: F03580376S19/2019©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: In bio-medical image processing, brain image segmentation is an aggressive concept in present days. Disorders of brain mainly requires accurate tissue extraction and classification of magnetic resonance (MR) medical brain images, which is very effective and important to detect different types of tumors, and necrotic tissue classification and segmentation. To handle brain image segmentation, mathematical tools like fuzzy sets, rough sets and soft sets are used to define uncertainty and vagueness of brain images. Accurate and effective segmentation and detection of tumor on brain image is still a challenging task in medical brain images with respect to reduction of noise, smoothness of image and accuracy for segmentation of medical brain images and other parameters. We propose and introduce a Novel Brain Segmentation approach based on neutrosophic soft sets is introduced to explore uncertainties relates to white, grey and cerebro spinal fluid matters for the detection tumor from MR brain image with respect to bias field estimation and co-relation based on decision making. Our proposed approach consist Pareto Optimization algorithm to support neutrosophic soft sets approximations for the optimal kernel parameters (like kernel functions). These approximations are free to define weight parameters and average, median, weight filters and less complexity compared to existing algorithms. Our experimental results show effective performance of proposed approach with respect to segmentation accuracy, time and jacquard parameters compared to existing algorithms.
Keywords: Segmentation of Brain Image, Fuzzy C-Means, Intuitionistic Neutrosophic Soft Sets, Rough Sets, Magnetic Resonance.
Scope of the Article: Clustering