Brain Tumor Segmentation Using Res-Unet

Accurate identification and segmentation of brain tumour is essential in medical imaging, as it plays a vital role in diagnosis, proper medication and improving patient outcomes. Traditional manual segmentation methods are time-consuming, labor-intensive, and prone to variability. Recently, deep learning techniques, particularly convolutional neural networks (CNNs), have shown promising results in medical image segmentation. This study proposes a novel brain tumor detection and segmentation framework using ResUNet, a hybrid architecture combining U-Net and ResNet, to leverage the advantages of both models. The framework is designed to handle both monomodal and multimodal MRI datasets, incorporating data augmentation and preprocessing techniques to improve model generalizability and robustness. Experimental results on the LGG MRI segmentation dataset demonstrate the superiority of the proposed framework, achieving high accuracy in tumor detection and segmentation, outperforming other contemporary segmentation techniques. Initial results demonstrate high accuracy of 93.75% in detecting and segmenting tumor regions, highlighting the potential of this model as a reliable, automated tool for brain tumor analysis in clinical settings. This study highlights the potential of deep learning in advancing healthcare, enabling accurate and efficient brain tumor diagnosis and treatment.