解决医学影像中数据分布偏移与类别不平衡问题，提升胚胎发育评估准确性
Addressing Domain Shift via Imbalance-Aware Domain Adaptation in Embryo Development Assessment
- 通过类别感知注意力机制实现自适应特征学习
- 动态加权对齐域间差异，最高提升25.19%准确率
- 适合低质量成像环境，适合临床部署的医疗模型研发
医学影像中的深度学习模型面临双重挑战：域偏移导致模型在不同采集环境下性能下降，以及某些疾病类型天然样本稀少造成的类别不平衡。本文提出不平衡感知域自适应（IADA）框架，通过三个核心组件协同应对：(1) 基于类别特异性注意力的自适应特征学习；(2) 动态加权的平衡域对齐；(3) 自适应阈值优化。理论分析提供了收敛性保证与复杂度边界。在四种成像模态的胚胎发育评估任务上，IADA显著优于现有方法，最高准确率提升25.19%，且各类别表现均衡。在低质量成像条件下，AUC最高提升12.56%，展现出强鲁棒性。结果表明，IADA具备构建可靠、公平的跨场景医疗影像系统潜力。代码已开源：https://github.com/yinghemedical/imbalance-aware_domain_adaptation。
原文摘要 · Abstract (English)
Deep learning models in medical imaging face dual challenges: domain shift, where models perform poorly when deployed in settings different from their training environment, and class imbalance, where certain disease conditions are naturally underrepresented. We present Imbalance-Aware Domain Adaptation (IADA), a novel framework that simultaneously tackles both challenges through three key components: (1) adaptive feature learning with class-specific attention mechanisms, (2) balanced domain alignment with dynamic weighting, and (3) adaptive threshold optimization. Our theoretical analysis establishes convergence guarantees and complexity bounds. Through extensive experiments on embryo development assessment across four imaging modalities, IADA demonstrates significant improvements over existing methods, achieving up to 25.19\% higher accuracy while maintaining balanced performance across classes. In challenging scenarios with low-quality imaging systems, IADA shows robust generalization with AUC improvements of up to 12.56\%. These results demonstrate IADA's potential for developing reliable and equitable medical imaging systems for diverse clinical settings. The code is made public available at \url{https://github.com/yinghemedical/imbalance-aware_domain_adaptation}
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。