A Kernel-Based Intuitionistic Fuzzy C-Means Clustering Using a DNA Genetic Algorithm for Magnetic Resonance Image Segmentation

• Published in: Entropy (Basel, Switzerland) Vol. 19; no. 11; p. 578
• Main Authors: Zang, Wenke, Zhang, Weining, Zhang, Wenqian, Liu, Xiyu
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 01.11.2017
• Subjects: Clustering; Computational efficiency; Computing time; Datasets; Deoxyribonucleic acid; DNA; DNA genetic algorithm; fuzzy C-means; Genetic algorithms; Image segmentation; images segmentation; intuitionistic fuzzy entropy; Magnetic resonance imaging; MRI; Solution space
• ISSN: 1099-4300; 1099-4300
• DOI: 10.3390/e19110578

MRI segmentation is critically important for clinical study and diagnosis. Existing methods based on soft clustering have several drawbacks, including low accuracy in the presence of image noise and artifacts, and high computational cost. In this paper, we introduce a new formulation of the MRI segmentation problem as a kernel-based intuitionistic fuzzy C-means (KIFCM) clustering problem and propose a new DNA-based genetic algorithm to obtain the optimal KIFCM clustering. While this algorithm searches the solution space for the optimal model parameters, it also obtains the optimal clustering, therefore the optimal MRI segmentation. We perform empirical study by comparing our method with six state-of-the-art soft clustering methods using a set of UCI (University of California, Irvine) datasets and a set of synthetic and clinic MRI datasets. The preliminary results show that our method outperforms other methods in both the clustering metrics and the computational efficiency.