We present a solution for the MIDOG 2025 Challenge Track~2, addressing binary classification of normal mitotic figures (NMFs) versus atypical mitotic figures (AMFs). The approach leverages pathology-specific foundation model H-optimus-0, selected based on recent cross-domain generalization benchmarks and our empirical testing, with Low-Rank Adaptation (LoRA) fine-tuning and MixUp augmentation. Implementation includes soft labels based on multi-expert consensus, hard negative mining, and adaptive focal loss, metric learning and domain adaptation. The method demonstrates both the promise and challenges of applying foundation models to this complex classification task, achieving reasonable performance in the preliminary evaluation phase.
翻译:我们提出针对MIDOG 2025挑战赛第二赛道的解决方案,致力于正常有丝分裂象(NMFs)与非典型有丝分裂象(AMFs)的二分类任务。该方法利用病理学专用基础模型H-optimus-0——该模型基于近期跨领域泛化基准测试及我们的实证评估而选定——并采用低秩自适应(LoRA)微调与MixUp数据增强技术。实施方案包含基于多专家共识的软标签标注、难负样本挖掘、自适应焦点损失函数、度量学习及领域自适应策略。该方法在初步评估阶段取得了合理性能,既展示了基础模型在此复杂分类任务中的应用潜力,也揭示了其面临的挑战。