Exploring Graph-Based Algorithms for Accurate Brain Tumor Segmentation

Accurate segmentation of brain tumours from Magnetic Resonance Imaging (MRI) is crucial for effective diagnosis and treatment planning, as variations in tumour size, shape, and location present significant challenges. This paper explores the potential of graph-based algorithms—Watershed, Graph Cut, GrabCut, and Random Walk—for precise brain tumour segmentation from MRI images. Utilizing the BRATS 2021 dataset, the Graph Cut algorithm achieved the best performance with a Dice Similarity Coefficient (DSC) of 0.8578, Jaccard Index (JI) of 0.9575, sensitivity of 0.9651, and specificity of 0.9998. Our findings demonstrate the effectiveness of graph-based methods, with Graph Cut excelling in handling noise and weak boundaries, providing high-quality segmentation. Also, this research highlights the potential of graph theory in medical image processing and enhancing clinical decision-making in brain tumour management.