当噪声遇上长尾：特征-阈值双重校准用于稳健伪标签
When Noise Meets Long-Tail: Feature-Threshold Dual Calibration for Robust Pseudo-Labeling
- Waseda University（早稻田大学）
- Dalian University of Technology（大连理工大学）
- The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对语义分割中噪声与长尾分布共存导致伪标签退化的问题，提出FTC-Seg框架，通过特征级正交原型重建和阈值级自适应校准，在四个基准上显著提升尾部类别性能。
AI中文摘要：
伪标签已成为语义分割中从无标注数据学习的关键技术。然而，在强成像噪声与长尾类分布同时出现的真实场景中，其有效性急剧下降。我们将这一失败归因于伪标签退化的恶性循环。成像噪声使前景与背景特征纠缠，降低所有类别的预测置信度，而长尾分布则使尾部类别训练样本更少，固有置信度更低。在固定的高阈值过滤下，这些尾部类别预测被系统性滤除，因此它们无法从无标注数据中获得监督，特征在后续迭代中持续退化。关键的是，噪声与长尾并非独立障碍，而是相互放大的，单独解决任一问题均不足。为打破这一循环，我们提出FTC-Seg，一种基于标准教师-学生框架的特征-阈值双重校准框架。在特征层面，正交原型重建（OPR）利用一组可学习的正交原型对像素级特征进行残差净化，扩大弱前景目标与噪声背景之间的间隔。在阈值层面，自适应阈值校准（ATC）根据学习难度和预测分布偏差动态调整类别特定阈值，将尾部类别的低置信度伪标签从系统性排除中拯救出来。在涵盖三种不同噪声模态的四个公开基准上的大量实验表明，FTC-Seg在性能上优于现有最先进方法，尤其在尾部类别上取得了显著提升。我们的结果证实，在复合噪声与类别不平衡下，联合校准特征和阈值对于稳健伪标签至关重要。
英文摘要：
Pseudo-labeling has become a cornerstone of learning from unlabeled data in semantic segmentation. Yet its effectiveness drops sharply in real-world scenarios where strong imaging noise and long-tailed class distributions occur together. We trace this failure to a vicious cycle of pseudo-label degradation. Imaging noise entangles foreground and background features, lowering prediction confidence across all classes, while long-tailed distributions leave tail classes with far fewer training samples and inherently lower confidence. Under fixed high-threshold filtering, these tail-class predictions are systematically filtered out, so they receive no supervision from unlabeled data and thus features keep degrading in subsequent iterations. Critically, noise and long-tail are not independent obstacles but mutually amplifying ones, and addressing either alone is insufficient. To break this cycle, we propose FTC-Seg, a Feature-Threshold dual-Calibration framework built on a standard teacher-student framework. At the feature level, Orthogonal Prototype Reconstruction (OPR) uses a set of learnable orthogonal prototypes to residually purify pixel-wise features, widening the margin between weak foreground targets and noisy backgrounds. At the threshold level, Adaptive Threshold Calibration (ATC) dynamically adjusts class-specific thresholds based on learning difficulty and prediction-distribution bias, rescuing low-confidence pseudo-labels of tail classes from systematic exclusion. Extensive experiments on four public benchmarks spanning three distinct noise modalities show that FTC-Seg achieves strong performance against state-of-the-art methods, with particularly substantial gains on tail classes. Our results establish that jointly calibrating features and thresholds is essential for robust pseudo-labeling under compounded noise and class imbalance.
