发表机构
Institute of Photogrammetry and Remote Sensing (IPF), Karlsruhe Institute of Technology (KIT)（卡尔斯鲁厄理工学院摄影测量与遥感研究所）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文首次系统评估四类不确定性量化方法在语义分割基础模型上的表现，揭示预测性能、可靠性与计算成本间的权衡，为现实应用需联合优化相关指标指明方向。
AI 中文摘要
基础模型正不断突破曾被认为不可能的局限，实现了前所未有的精度与跨域泛化能力。然而，其缺乏可解释性、易过度自信以及对现实域偏移敏感的特性，对安全及关键任务应用构成重大挑战。不确定性量化（UQ）为解决这些问题提供了原则性方法，但其与分割基础模型的整合尚未得到探索。本文首次对应用于语义分割基础模型的UQ方法开展系统评估：我们在预训练的SAM2编码器之上微调轻量DPT解码器，构建简单却具竞争力的基线；在Cityscapes、NYUv2及两个具有挑战性的域外场景中，对四种代表性UQ方法——蒙特卡洛dropout、深度子集成、测试时增强、证据深度学习——进行基准测试。我们的分析对比了分割精度、校准度、不确定性质量与推理时间，揭示了预测性能、可靠性与计算成本间的明显权衡。这些结果凸显了感知不确定性的基础模型的潜力与当前局限，表明未来需开展联合优化精度、鲁棒性与效率的研究以实现现实部署。
英文摘要
Foundation models are increasingly breaking what seemed to be impossible not long ago by enabling unprecedented accuracy and cross-domain generalization. Yet their lack of interpretability, tendency to be overconfident, and sensitivity to real-world domain shifts pose critical challenges for safety- and mission-critical applications. Uncertainty quantification (UQ) offers a principled way to address these issues, but its integration into segmentation foundation models has yet to be explored. In this paper we present the first systematic evaluation of UQ methods applied to a foundation model for semantic segmentation. We fine-tune a lightweight DPT decoder on top of the pretrained SAM2 encoder to establish a simple yet competitive baseline and benchmark four representative UQ approaches - Monte Carlo Dropout, Deep Sub-Ensemble, Test-Time Augmentation, and Evidential Deep Learning - across Cityscapes, NYUv2, and two challenging out-of-domain settings. Our analysis compares segmentation accuracy, calibration, uncertainty quality, and inference time, revealing clear trade-offs between predictive performance, reliability, and computational cost. These results highlight both the promise and the current limitations of uncertainty-aware foundation models, pointing to the need for future work that jointly optimizes accuracy, robustness, and efficiency for real-world deployment.
CommentsAccepted for publication in the ISPRS Annals (ISPRS Congress 2026, Toronto, Oral Presentation)
