Segmentation of Fetal Images using Kernel Fuzzy C Means Clustering with Whale Optimisation Algorithm
S. K. Rajalakshmi1, S. Sivagamasundari2

1S.K. Rajalakshmi, Research Scholar, Department of E & I, Annamalai University, Tamilnadu, India.
2Dr. S. Sivagamasundari, Professor, Department of E & I, Annamalai University, Tamilnadu, India.
Manuscript received on 25 August 2019. | Revised Manuscript received on 09 September 2019. | Manuscript published on 30 September 2019. | PP: 4122-4125 | Volume-8 Issue-11, September 2019. | Retrieval Number: K14820981119/2019©BEIESP | DOI: 10.35940/ijitee.K1482.0881119
Open Access | Ethics and 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: Image segmentation is considered as a critical process in medical imaging that are facilitated using automated computation. The segmentation process partitions the images into subsets based on its location or intensity. However, segmentation of fetal images faces poor segmentation due to the presence of noise and poor spatial intensities. In this paper, the study decomposes the fetal image into several parts for the purpose of segmentation and then performs the change in representations. The segmentation process is improved in this method using Kernel Fuzzy C Means (KFCM) based Whale Optimisation Algorithm (WOA). The segmentation process uses modified KFCM, where the centroid values are estimated using WOA. The segmentation method segments the input fetal image into appropriate regions using KFCM-WOA. The simulation result shows that the proposed method attains improved performance than other kernel based methods. The results of the performance metrics shows that the proposed method attains a sensitivity of 99.8273%, specificity of 99.7350%, accuracy of 99.9385%, positive predictive value (PPV) of 99.3964, Negative Predictive value (NPV) of 0.3805, Dice Coefficient of 48.5518, Rand Index (RI) of 0.9983 and Global consistency error (GCE) of 0.0460, which are higher than other kernel based methods.
Keywords: WOA, Segmentation, KFCM, Centroid estimation.
Scope of the Article: Data Analytics Modelling and Algorithms