BUFNet: Boundary-aware and uncertainty-driven multi-modal fusion network for MR brain tumor segmentation.

Abstract

Brain tumor segmentation plays a critical role in the diagnosis and treatment planning of brain tumors. However, achieving accurate segmentation is challenging due to the complex boundaries between different tumor sub-regions. Additionally, many existing methods produce deterministic segmentation results without addressing prediction uncertainty, limiting their reliability and interpretability in clinical applications. To tackle these challenges, this paper proposes a novel Boundary-aware and Uncertainty-driven multi-modal Fusion Network (BUFNet) for MR brain tumor segmentation. Specifically, a boundary-aware mechanism is proposed to extract tumor boundary information, and guide the network by leveraging this information for better discrimination of tumor sub-regions. Furthermore, an effective multi-modal fusion method is proposed to integrate complementary information from multiple MR modalities. To further reduce uncertainty, a novel uncertainty-based segmentation loss function is proposed to improve segmentation performance. Additionally, to enhance clinical interpretation and decision-making, uncertainty quantification is incorporated to provide confidence measures for segmentation results. Experimental results demonstrate the effectiveness of the proposed method, showing superior performance compared to state-of-the-art methods.