An Optimized Clustering Approach using Tree Seed Algorithm for the Brain MRI Images Segmentation

Clustering algorithms are widely used for segmenting medical images. However, these techniques can be challenging to perform, especially when working with magnetic resonance images (MRI) of the brain. The complexity of the brain tissue's anatomical structure, the inhomogeneity of the pixel intensity in these images, and the effects of partial volume and noise can make clustering difficult, leading the algorithm to fall into local minima. To address this issue, it is recommended to improve clustering algorithms by using optimization techniques to achieve better results. In this study, we propose a developed clustering algorithm and optimize it using a tree seed algorithm (TSA) for segmenting brain MRI images. The algorithms are tested on real brain image datasets, and the experimental results show that our proposed and optimized methods yield satisfactory results according to the Davies-Bouldin Index (DBI), compared to the fuzzy c-mean (FCM) algorithm.