HyFormer-Net：一种用于超声图像乳腺病灶分割与分类的具有可解释多尺度融合的协同CNN-Transformer
HyFormer-Net: A Synergistic CNN-Transformer with Interpretable Multi-Scale Fusion for Breast Lesion Segmentation and Classification in Ultrasound Images
AI总结：
提出混合架构HyFormer-Net，通过多尺度融合与双管道可解释性同时进行乳腺超声病灶分割与分类，在BUSI数据集上表现优异，并首次验证了其跨数据集渐进式微调的泛化能力。
摘要：
用于乳腺癌诊断的B型超声面临挑战：斑点噪声、操作者依赖性以及边界不清晰。现有的深度学习存在单任务学习、架构限制（CNN缺乏全局上下文，Transformer缺乏局部特征）以及黑盒决策等问题。这些缺陷阻碍了其临床应用。我们提出了HyFormer-Net，一种混合CNN-Transformer，用于同时进行分割和分类并具有内在可解释性。其双分支编码器通过多尺度分层融合块集成了EfficientNet-B3和Swin Transformer。注意力门控解码器提供了精度和可解释性。我们引入了双管道可解释性：（1）具有定量IoU验证（平均值：0.86）的内在注意力验证，以及（2）用于分类推理的Grad-CAM。在BUSI数据集上，HyFormer-Net实现了0.761 +/- 0.072的Dice Score和93.2%的准确率，优于U-Net、Attention U-Net和TransUNet。92.1 +/- 2.2%的Malignant Recall确保了极少的假阴性。集成建模产生了卓越的Dice 90.2%、准确率99.5%以及完美的100% Malignant Recall，消除了假阴性。消融研究证实多尺度融合贡献了+16.8%的Dice，注意力门增加了+5.9%。至关重要的是，我们进行了混合CNN-Transformers在乳腺超声中的首次跨数据集泛化研究。零样本迁移失败（Dice：0.058），证实了域偏移。然而，仅使用10%目标域数据（68张图像）进行渐进式微调即可恢复92.5%的性能。使用50%数据，我们的模型达到了77.3%的Dice，超过了源域性能（76.1%）并展示了真正的泛化能力。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
