发表机构
Graduate School of Information Sciences, Tohoku University; Department of Ophthalmology, Graduate School of Medicine, Tohoku University; Institute of Ophthalmology, University College London; NIHR Biomedical Research Centre at Moorfields Eye Hospital NHS Foundation Trust; UCL Hawkes Institute, University College London; Department of Computer Science, University College London（东北大学信息科学研究生院; 东北大学医学研究生院眼科系; 伦敦大学学院眼科研究所; 摩尔菲尔德眼科医院NHS基金会信托基金NIHR生物医学研究中心; 伦敦大学学院UCL霍克斯研究所; 伦敦大学学院计算机科学系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对OCT图像体积分割需大量标注的问题，提出PC-Seg框架，通过单个二维模型学习跨视图一致性生成伪标签，融入三维模型并协同训练，在仅用少量训练数据标签时，精度与全监督相当且优于现有方法。
AI 中文摘要
光学相干断层扫描（OCT）图像的体积分割对眼部疾病诊断至关重要，但需要大量体素级标注。半监督学习可降低标注成本，但多数现有方法逐片处理数据，未利用固有三维空间上下文。我们提出PC-Seg，一种渐进式跨视图一致性框架，从稀疏二维标注学习高精度三维分割模型。它用单个二维模型从标准B扫描和正交切片学习跨视图一致性，生成可靠体积伪标签，再融入三维模型，经协同训练阶段，二维和三维模型通过集成伪标签相互优化。在MSHC和Duke DME数据集上的实验表明，PC-Seg在仅使用约0.7%训练数据标签时，精度与全监督学习相当，优于现有半监督和视网膜层分割方法。
英文摘要
Volumetric segmentation of optical coherence tomography (OCT) images is essential for diagnosing ocular diseases but requires labor-intensive voxel-wise annotations. While semi-supervised learning (SSL) can reduce annotation costs, most existing methods process data slice by slice and fail to exploit the inherent 3D spatial context. We propose PC-Seg, a progressive cross-view consistency framework that learns high-accuracy 3D segmentation models from sparse 2D annotations. Unlike conventional multi-view approaches, PC-Seg uses a single 2D model to learn cross-view consistency from standard B-scans and orthogonal slices, thereby generating reliable volumetric pseudo-labels. These pseudo-labels are then distilled into a 3D model, followed by a co-training stage in which the 2D and 3D models mutually refine each other through ensemble pseudo-labeling. Experiments on the MSHC and Duke DME datasets demonstrate that PC-Seg achieves accuracy comparable to fully supervised learning while using labels for only about 0.7% of the training data, outperforming state-of-the-art semi-supervised and retinal layer segmentation methods. Our code is publicly available at https://github.com/gsisaoki/pc-seg-official.
CommentsAccepted at MICCAI 2026
