Brain Segmentation using Adaptive Thresholding, K-Means Clustering and Mathematical Morphology in MRI Data

Abstract

Nowadays, many methods have been applied for brain segmentation on MRI data. This paper proposes a new method for brain segmentation using Adaptive Thresholding, K-Means Clustering, and Morphological Mathematics in MRI data. The adaptive threshold was chosen because the adaptive threshold method will vary across images to suit various lighting conditions and background changes. We segment the corpus callosum. This experiment shows that with the Adaptive Thresholding, K-Means Clustering, and Mathematical Morphology to segment the corpus callosum produces the highest Dice Similarity Coefficient (DSC) value of 0.757.