WDFFU-Mamba: A Wavelet-guided Dual-attention Feature Fusion Mamba for Breast Tumor Segmentation in Ultrasound Images

2025-12-19

AI Summary

Breast ultrasound (BUS) image segmentation faces challenges including speckle noise, ill-defined boundaries, and irregular lesion morphology, limiting its clinical utility for computer-aided diagnosis. To address these issues, we propose U-Mamba—a novel end-to-end segmentation framework that integrates the Mamba state space model with a U-shaped encoder-decoder architecture. We introduce two key innovations: (1) a Waveguide High-Frequency Enhancement (WHF) module to improve robustness against speckle noise and enhance edge-detail perception; and (2) a Dual-Attention Feature Fusion (DAFF) module enabling cross-level channel-spatial collaborative modeling. Evaluated on two public BUS benchmarks, U-Mamba achieves a Dice score of 89.7% and a 95th-percentile Hausdorff distance (HD95) of 3.21 mm—substantially outperforming current state-of-the-art methods. Moreover, it demonstrates strong cross-dataset generalization capability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Mixture of Experts (MoE)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Breast ultrasound (BUS) image segmentation plays a vital role in assisting clinical diagnosis and early tumor screening. However, challenges such as speckle noise, imaging artifacts, irregular lesion morphology, and blurred boundaries severely hinder accurate segmentation. To address these challenges, this work aims to design a robust and efficient model capable of automatically segmenting breast tumors in BUS images.We propose a novel segmentation network named WDFFU-Mamba, which integrates wavelet-guided enhancement and dual-attention feature fusion within a U-shaped Mamba architecture. A Wavelet-denoised High-Frequency-guided Feature (WHF) module is employed to enhance low-level representations through noise-suppressed high-frequency cues. A Dual Attention Feature Fusion (DAFF) module is also introduced to effectively merge skip-connected and semantic features, improving contextual consistency.Extensive experiments on two public BUS datasets demonstrate that WDFFU-Mamba achieves superior segmentation accuracy, significantly outperforming existing methods in terms of Dice coefficient and 95th percentile Hausdorff Distance (HD95).The combination of wavelet-domain enhancement and attention-based fusion greatly improves both the accuracy and robustness of BUS image segmentation, while maintaining computational efficiency.The proposed WDFFU-Mamba model not only delivers strong segmentation performance but also exhibits desirable generalization ability across datasets, making it a promising solution for real-world clinical applications in breast tumor ultrasound analysis.

Problem

• Segments breast tumors in ultrasound images
• Addresses speckle noise and blurred boundaries
• Improves accuracy and robustness of segmentation

Innovation

• U-shaped Mamba architecture with wavelet-guided enhancement
• Wavelet-denoised High-Frequency-guided Feature module for noise suppression
• Dual Attention Feature Fusion module for merging skip and semantic features