SEFF-Net: A Hybrid Feature Fusion Network for Accurate Segmentation of Breast Ultrasound Images

Breast ultrasound imaging plays a crucial role in early breast cancer screening and diagnosis due to its noninvasive nature and cost-effectiveness. However, accurate lesion segmentation remains challenging because of severe speckle noise, low contrast, and blurred tumor boundaries. To address these issues, this paper proposes SEFF-Net, a novel edge-aware feature fusion network with a U-shaped encoder–decoder architecture to capture multi-level semantic representations for breast ultrasound image segmentation task. To enhance boundary perception, a Self-learning Edge Enhancement Module is embedded in the shallow encoding stages, while a Spatial Feature Fusion Module is introduced to effectively integrate multi-scale features by leveraging spatial context, thereby achieving a better balance between low-level spatial details and high-level semantic information.To further alleviate the class imbalance between foreground and background regions and improve boundary learning, a novel joint loss function is designed by combining region-based consistency constraints with boundary-sensitive supervision. This optimization strategy reinforces contour awareness while maintaining overall segmentation accuracy. Experimental results demonstrate that SEFF-Net consistently outperforms state-of-the-art segmentation methods across multiple evaluation metrics, including Dice coefficient, IoU, and boundary-related measures. Overall, SEFF-Net provides an effective and reliable solution for accurate breast ultrasound image segmentation, showing promising potential for clinical computer-aided diagnosis systems.