arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~

MSA-DCNN:一种用于医学图像分类的数据高效多尺度可变形卷积神经网络

MSA-DCNN: A Data-Efficient Multi-Scale Attention Deformable CNN for Medical Image Classification

Hamza Hussaini, Shahana Bano, Eyad Elyan, Carlos Francisco Moreno-García

arXiv 2607.06083首次发表:更新:

发表机构

Robert Gordon University(罗伯特·戈登大学)

机构由 AI 辅助整理,请以论文原文为准。

AI 中文总结

针对现有医学图像分类方法难题,提出MSA-DCNN框架,引入自适应多尺度采样等技术,在统一优化方案中学习跨尺度融合等。实验表明其在多方面性能优于基线且参数少,消融实验证实组件贡献,为数据高效医学图像分类提供基础。

AI 中文摘要

现有深度学习方法在医学图像分类中表现良好,但由于固定采样和数据密集型训练,在处理多尺度形态和有限标注方面存在困难。现有方法孤立地应对这些挑战:基于可变形卷积网络(DCN)的模型提供自适应采样,但缺乏明确的多尺度注意力融合和标签高效正则化;多尺度架构通常依赖静态融合;半监督方法针对标签稀缺问题,而没有联合建模自适应跨尺度表示。我们提出了MSA-DCNN,这是一个尺度一致的可变形注意力学习框架,在统一优化方案中引入了自适应多尺度采样、尺度内显著性细化、学习到的跨尺度融合和辅助自蒸馏,有可能推广到结构异质的解剖结构。我们在三个公共基准和一个白血病外部保留集上进行评估。在分布偏移和标签稀缺情况下,MSA-DCNN在准确率、F1值和AUC(二元)方面,相对于视觉Transformer(ViT)基线、卷积神经网络(CNN)基线和一个医学图像计算方法国际会议(MICCAI)半监督基线,展现出有竞争力且通常更好的性能,同时使用更少参数。消融实验证实了各组件的互补贡献,支持MSA-DCNN作为数据高效医学图像分类的实用基础。

英文摘要

Existing deep learning methods perform well in medical image classification but struggle with multi-scale morphology and limited annotations due to fixed sampling and data-hungry training. Existing approaches address these challenges in isolation: DCN-based models provide adaptive sampling but lack explicit multi-scale attention fusion and label-efficient regularisation; multi-scale architectures typically rely on static fusion; and semi-supervised methods target label scarcity without jointly modelling adaptive cross-scale representations. We propose MSA-DCNN, a scale-consistent deformable attention learning framework that introduces adaptive multi-scale sampling, within-scale saliency refinement, learned cross-scale fusion, and auxiliary self-distillation within a unified optimisation scheme, with potential to generalise to structurally heterogeneous anatomy. We evaluate on three public benchmarks and an external hold-out set for leukaemia. MSA-DCNN demonstrates competitive and often better performance against ViT baselines, CNN baselines, and a MICCAI semi-supervised baseline under distribution shift and label scarcity in accuracy, F1, and AUC (binary), while using fewer parameters. Ablations confirm complementary component contributions, supporting MSA-DCNN as a practical foundation for data-efficient medical image classification.

论文原文

arXiv 摘要页 · PDF 原文 · HTML 原文

↑