用干净正样本提升大模型半监督分割，无需复杂阈值
PixCon: Clean-Positive Contrastive Learning for Foundation-Model Semi-Supervised Segmentation
- 构建仅含正确标签像素的类别记忆库，确保正样本无污染
- 在Pascal VOC等数据集上比基线提升0.2 mIoU，均值达87.90
- 适合使用大模型做半监督分割的研究者，降低调参成本
半监督语义分割长期关注伪标签可信度问题，传统方法依赖严格置信度过滤。当采用DINOv2教师模型时，简单阈值即可保留98%清洁的伪标签，此时性能瓶颈转向嵌入空间的类别结构设计。本文提出PixCon——一种基于干净正样本的像素对比学习框架。其为每类维护一个记忆库，仅允许学生模型已正确分类的标注像素进入，从构造上保证正样本无污染（ρ_F=0），不同于以往依赖置信度筛选的ReCo、U²PL等方法。该方法为单分支架构，不增加推理参数，无需特定阈值。首次对监督版InfoNCE梯度进行一阶分析，揭示污染危害：错误正样本项随ρ_F/(1−ρ_F)增长，实测ρ_F在Pascal为0.018，ADE20K为0.106。在Pascal VOC、Cityscapes和ADE20K上，PixCon在计算量匹配的一次切换协议下，与强基线UniMatch V2持平或超越，所有Pascal-1/8种子均提升约+0.2 mIoU，三种子平均达87.90，超过已发表的UniMatch V2-B结果。由于大模型教师下污染本就稀少，分析表明ρ_F=0的保障主要增强鲁棒性，而精度提升源于更清晰的正样本监督，因此干净正样本对比是大模型半监督分割的稳健且低成本默认方案。
原文摘要 · Abstract (English)
Semi-supervised semantic segmentation (SSSS) has long turned on one question, which pseudo-labels to trust, and answered it with ever more careful confidence filtering. Foundation backbones change the regime: with a DINOv2 teacher a strict threshold already retains a measured 98%-clean pseudo-label set, so the accuracy that remains lives not in the filter but in how the embedding space is structured by class. We propose PixCon, a clean-positive pixel-contrastive framework. PixCon maintains a per-class memory bank that admits only labeled pixels the student already classifies correctly, guaranteeing a contamination-free positive set ($ρ_F=0$) by construction, unlike prior contrastive SSSS banks (ReCo, U$^2$PL) built from confidence-filtered pseudo-labels. It is a single branch over a consistency backbone, adds no inference-time parameters, and needs no bank-specific threshold. A first-order analysis of the supervised-InfoNCE gradient explains why contamination hurts: its false-positive term scales as $ρ_F/(1-ρ_F)$, which we measure (0.018 on Pascal, 0.106 on ADE20K) rather than assume. Across Pascal VOC, Cityscapes, and ADE20K, PixCon matches or improves a strong DINOv2-based UniMatch V2 baseline in a compute-matched one-switch protocol: it improves every Pascal-1/8 seed (a per-seed gain of about +0.2 mIoU) and its three-seed mean reaches 87.90, the published UniMatch V2-B figure. Because contamination is already rare under foundation-model teachers, our analysis indicates the $ρ_F=0$ guarantee acts chiefly as robustness as teachers weaken, while the accuracy gain comes from cleaner positive supervision, making clean-positive contrast a robust, low-cost default for foundation-model SSSS.
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。