GaMNet: A Hybrid Network with Gabor Fusion and NMamba for Efficient 3D Glioma Segmentation

This chapter introduces GaMNet, a hybrid network designed for efficient 3D glioma segmentation, focusing on key topics such as the integration of Gabor filters and NMamba, the architecture of the hybrid network, experimental results, and comparisons with state-of-the-art methods. The study addresses the challenges of tumor heterogeneity and shape irregularity, proposing a solution that combines local and global feature extraction. The network's performance is evaluated on the BraTS2023 and FLARE2022 datasets, demonstrating superior accuracy and efficiency. The results show that GaMNet outperforms existing methods in terms of Dice score, Hausdorff Distance 95%, False Negative Rate, and False Positive Rate, making it a promising tool for accurate and reliable glioma segmentation. The chapter also includes an ablation study to highlight the contributions of each component and an efficiency study to assess the network's balance between complexity and performance. Overall, the chapter provides a comprehensive overview of the proposed method and its advantages, offering valuable insights for professionals in the field of medical image analysis.