Proceedings of the International conference on Applied Research in Engineering,Science and Technology
Year: 2018 | Page No:121-133
Development of an Enhanced Fuzzy C-Means algorithm for Medical Image Segmentation using Ant Colony Optimization
Ajala F.A, Ojebamigbe V.I and Fenwa O.D
Modified Fuzzy C-Means (MFCM) algorithm, which is a combination of fuzzy c means and k means, for medical image segmentation, suffers from high computational time and noise which lead to high memory consumption and low accuracy. Therefore, this paper presents an Enhanced MFCM algorithm (ACOMFCM), which has better accuracy and low memory consumption for medical image segmentation. Thirty medical images used in this work were pre-processed using Gaussian filtering method. ACOMFCM algorithm was developed using Ant Colony Optimization techniques to minimise the Euclidean distance between the data point and centre coordinate in the K-means algorithm characterizing the existing MFCM. Segmentation of the thirty images were carried out using the MFCM and ACOMFCM algorithms in Matrix Laboratory 7.1 (R0011a) environment. Performance of the MFCM and the ACOMFCM was evaluated using segmentation accuracy, segmentation time and memory consumption.
The average result of MFCM algorithm for thirty images yielded segmentation accuracy, segmentation time and memory consumption of 7117992.29, 22.484s and 494161920bit, respectively while average result of the ACOMFCM algorithm for thirty images used yielded segmentation accuracy, segmentation time and memory consumption of 10590135.79, 6.649s and 502960128 bit, respectively.
Keywords: MFCM, ACOMFCM, Segmentation, Accuracy, Memory Consumption.