EUnet++: Enhanced Unet++ Architecture Incorporating 3DCNN and CBAM Modules for Brain Tumor Image Segmentation

Brain tumor segmentation plays a crucial role in medical imaging, aiding in accurate diagnosis and treatment planning. This paper proposes EUnet++, an enhanced Unet++ architecture incorporating 3D Convolutional Neural Networks (3DCNN) and Convolutional Block Attention Modules (CBAM) to improve segmentation accuracy. The 3DCNN module enhances the model’s ability to capture spatial features, while CBAM focuses attention on critical tumor regions for better segmentation performance. Through ablation studies, comparative experiments, and visualizations, the study show that EUnet++ outperforms traditional models in terms of segmentation precision, particularly in capturing complex tumor boundaries and fine details. These findings underscore the potential of EUnet++ in advancing brain tumor segmentation tasks in clinical applications.