Volumetric Brain Tumor Segmentation Using V-Net

Brain tumors are caused by rapid and uncontrolled cell growth that may pose a potential threat to life if not treated at an early stage. The primary difficulty in this research domain is the accurate segmentation of both tumor and non-tumor regions due to the intricate structure of the brain. The main challenge in detecting brain tumors arises from the irregularities in the shape and location of the tumor. Accurate segmentation of brain tumors in Magnetic Resonance Imaging (MRI) images is essential for clinical diagnosis and helps to make decisions for patient treatment. The research outlined in this study utilized the V-Net architecture, employing an encoder-decoder structure to perform volumetric brain tumor segmentation in MRI images. V-Net architecture is trained and evaluated on the BraTS2020 dataset employing different loss functions. The performance measures utilized to assess the obtained segmentation results in this research are the dice score, specificity, and sensitivity.