Brain MRI Segmentation for Lesion Detection Using Clustering with Fire-Fly Algorithm

This paper presents brain MRI segmentation for lesion detection using fire-fly based hard-clustering algorithm. First, MR images are denoised using median filter and denoised images are segmented using fire-fly based clustering algorithm. After segmentation, lesioned regions are extracted from segmented MR images. The performance of the proposed method is evaluated using quantitative measurement index. A comparative study is made with k-means and Fuzzy c-means algorithms. The experiment results demonstrate that the proposed method performs better than other two methods.