HADS-Net:A Hybrid Attention-Augmented Dual-Stream Network with Physics-Informed Augmentation for Breast Ultrasound Image Classification

2026-05-19

AI Summary

This study addresses the challenge of classifying benign, malignant, and normal tissues in breast ultrasound images, which is hindered by speckle noise, acoustic shadowing, and inter-class visual ambiguity. To tackle this, the authors propose a novel dual-stream network architecture that integrates a local stream specifically designed for lesion boundary refinement with a physics-informed global texture stream. Feature fusion and joint optimization are achieved through a cross-attention mechanism. Built upon an EfficientNet-B3 backbone, the model incorporates Sobel edge extraction, physics-aware data augmentation, and an adaptive class-weighted Focal Loss. Evaluated on the BUSI dataset, the method achieves 96.58% accuracy, a macro ROC-AUC of 0.9978, and a macro F1-score of 0.9654, with zero malignant lesions misclassified as normal—demonstrating substantially improved diagnostic reliability.

Abstract

Accurate classification of breast ultrasound images into benign, malignant, and normal categories is a critical clinical task complicated by speckle noise, acoustic shadowing, and inter-class visual ambiguity. Existing deep learning methods rely on single-stream architectures with generic augmentation that ignores ultrasound acquisition physics, and no prior method dedicates a stream to the lesion boundary features identified as the most diagnostically significant visual cue. We propose HADS-Net, a Hybrid Attention-Augmented Dual-Stream Network exploiting global texture and local boundary cues through two parallel pathways. Stream 1 applies physics-informed augmentation simulating speckle noise, acoustic shadowing, and gain variation before extracting features via pretrained EfficientNet-B3 projected to 512 dimensions. Stream 2 extracts Sobel edge maps processed by a lightweight CNN projected to the same 512-dimensional space. A cross-attention fusion module allows the texture stream to selectively query boundary features, producing a jointly optimised representation classified by an MLP trained with adaptive class-weighted focal loss. Five-fold stratified cross-validation with cosine annealing over 50 epochs is used, with the globally best checkpoint selected by lowest validation loss evaluated on a held-out test set. On the BUSI dataset, HADS-Net achieves 96.58% accuracy, macro ROC-AUC of 0.9978, macro F1 of 0.9654, and per-class F1-scores of 0.970, 0.951, and 0.976 for benign, malignant, and normal. No malignant lesion is misclassified as normal. These results confirm that modality-specific augmentation with cross-modal attention fusion is an effective strategy for ultrasound-based breast cancer diagnosis.

Problem

• breast ultrasound image classification
• speckle noise
• acoustic shadowing
• inter-class visual ambiguity
• lesion boundary features

Innovation

• dual-stream network
• physics-informed augmentation
• cross-attention fusion
• boundary-aware feature extraction
• ultrasound image classification