DA-Cal: 向语义分割中的跨域校准迈进
DA-Cal: Towards Cross-Domain Calibration in Semantic Segmentation
AI总结：
DA-Cal通过引入元温度网络和双层优化，提升语义分割中跨域校准效果，减少域差异并提高性能。
摘要：
尽管现有的无监督域适应（UDA）方法在语义分割中显著提升了目标域的性能，但它们往往忽视了网络校准质量，导致预测置信度与实际准确性之间的不一致--这在安全关键应用中是一个重大风险。我们的关键见解来自于观察到，在跨域场景中，当软伪标签替代硬伪标签时，性能会显著下降，这归因于校准不佳，尽管理论上完美校准的软伪标签等同于硬伪标签。基于这一发现，我们提出了DA-Cal，一个专门的跨域校准框架，将目标域校准转换为软伪标签优化。DA-Cal引入了一个元温度网络来生成像素级校准参数，并利用双层优化来建立软伪标签与UDA监督之间的关系，同时利用互补的域混合策略来防止过拟合并减少域差异。实验表明，DA-Cal可以无缝集成到现有的自训练框架中，在多个UDA分割基准上取得了显著成效，同时在不增加推理开销的情况下提高了目标域的校准效果。代码将被发布。
英文摘要：
While existing unsupervised domain adaptation (UDA) methods greatly enhance target domain performance in semantic segmentation, they often neglect network calibration quality, resulting in misalignment between prediction confidence and actual accuracy---a significant risk in safety-critical applications. Our key insight emerges from observing that performance degrades substantially when soft pseudo-labels replace hard pseudo-labels in cross-domain scenarios due to poor calibration, despite the theoretical equivalence of perfectly calibrated soft pseudo-labels to hard pseudo-labels. Based on this finding, we propose DA-Cal, a dedicated cross-domain calibration framework that transforms target domain calibration into soft pseudo-label optimization. DA-Cal introduces a Meta Temperature Network to generate pixel-level calibration parameters and employs bi-level optimization to establish the relationship between soft pseudo-labels and UDA supervision, while utilizing complementary domain-mixing strategies to prevent overfitting and reduce domain discrepancies. Experiments demonstrate that DA-Cal seamlessly integrates with existing self-training frameworks across multiple UDA segmentation benchmarks, significantly improving target domain calibration while delivering performance gains without inference overhead.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
