混合卷积与Transformer模型，实现乳腺超声图像病灶分割与分类的精准可解释分析
HyFormer-Net: A Synergistic CNN-Transformer with Interpretable Multi-Scale Fusion for Breast Lesion Segmentation and Classification in Ultrasound Images
- 双分支编码器融合EfficientNet-B3与Swin Transformer，多尺度层级融合提升特征表达
- 分割Dice达0.761，分类准确率93.2%，恶性病灶召回率达92.1%，且无漏诊
- 首次验证跨数据集泛化能力，少量微调即可恢复性能，适合临床部署
B型超声在乳腺癌诊断中面临斑点噪声、操作者依赖及边界模糊等挑战。现有深度学习方法存在单任务学习、结构局限（CNN缺乏全局上下文，Transformer缺少局部细节）和决策黑箱问题，阻碍临床应用。本文提出HyFormer-Net，一种兼顾分割与分类的混合CNN-Transformer架构，具备内在可解释性。其双分支编码器通过多尺度层级融合块结合EfficientNet-B3与Swin Transformer；注意力门控解码器提升精度与可解释性。引入双管道可解释性：(1) 内在注意力验证结合定量IoU（均值0.86），(2) Grad-CAM揭示分类依据。在BUSI数据集上，模型分割Dice为0.761±0.072，分类准确率93.2%，恶性病灶召回率92.1±2.2%，优于U-Net、Attention U-Net与TransUNet。集成模型表现更佳：Dice达90.2%，准确率99.5%，恶性召回100%。消融实验证明多尺度融合带来+16.8% Dice提升，注意力门控贡献+5.9%。首次开展混合模型在乳腺超声中的跨数据集泛化研究：零样本迁移失败（Dice: 0.058），但仅用10%目标域数据（68张图像）进行渐进式微调，即恢复92.5%性能；50%数据时达到77.3% Dice，超过源域性能（76.1%），证明真正泛化能力。
原文摘要 · Abstract (English)
B-mode ultrasound for breast cancer diagnosis faces challenges: speckle, operator dependency, and indistinct boundaries. Existing deep learning suffers from single-task learning, architectural constraints (CNNs lack global context, Transformers local features), and black-box decision-making. These gaps hinder clinical adoption. We propose HyFormer-Net, a hybrid CNN-Transformer for simultaneous segmentation and classification with intrinsic interpretability. Its dual-branch encoder integrates EfficientNet-B3 and Swin Transformer via multi-scale hierarchical fusion blocks. An attention-gated decoder provides precision and explainability. We introduce dual-pipeline interpretability: (1) intrinsic attention validation with quantitative IoU verification (mean: 0.86), and (2) Grad-CAM for classification reasoning. On the BUSI dataset, HyFormer-Net achieves Dice Score 0.761 +/- 0.072 and accuracy 93.2%, outperforming U-Net, Attention U-Net, and TransUNet. Malignant Recall of 92.1 +/- 2.2% ensures minimal false negatives. Ensemble modeling yields exceptional Dice 90.2%, accuracy 99.5%, and perfect 100% Malignant Recall, eliminating false negatives. Ablation studies confirm multi-scale fusion contributes +16.8% Dice and attention gates add +5.9%. Crucially, we conduct the first cross-dataset generalization study for hybrid CNN-Transformers in breast ultrasound. Zero-shot transfer fails (Dice: 0.058), confirming domain shift. However, progressive fine-tuning with only 10% target-domain data (68 images) recovers 92.5% performance. With 50% data, our model achieves 77.3% Dice, exceeding source-domain performance (76.1%) and demonstrating true generalization.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。