Abstract: Images in the field of medicine are of vital interest to medical practitioners because they help to diagnose the right disease. They are used to find the abnormalities in the working and status of the human system, based on which the correct treatment can be planned for the patients. But, the images may not be accurate and may contain noise which may have occurred due to various reasons. Therefore, it is necessary to improve the quality of the image. Segmentation helps to identify the occurrence of any abnormality in the human body from the MRI image. In the proposed system, the MRI brain image is preprocessed using Gaussian filtering to enhance the quality of the image by removing the noise. Segmentation is performed on the preprocessed image by clustering using the optimal Multi-Objective Adaptive Fuzzy C Means (MAFCM) algorithm which combines both FCM and Cuckoo search algorithms for identifying the important parts of the brain that help to identify the disease, example brain tumor. The multi-objective feature of the proposed algorithm leads to optimal results which is better than the existing techniques in terms of computation complexity and time complexity. The performance of the proposed system is compared with the existing Adaptive Fuzzy K Means (AFKM) and Optimally Enhanced Fuzzy K-C-Means (OEFKCM) segmentation methods in terms of the parameters such as structural similarity index measure, structural content, mean square error and peak signal to noise ratio and is found to exhibit better results in terms of efficient identification of the disease.
J. Mercy Geraldine, E. Kirubakaran and S. Sathiya Devi, 2016. MRI Brain Image Segmentation by Clustering Using an Optimal Multi-Objective Adaptive Fuzzy C Means (MAFCM) Algorithm. Asian Journal of Information Technology, 15: 54-64.