aiqbench analysis
This research addresses critical miscalibration issues in semantic segmentation by introducing translation-invariant calibrators and decision-preserving methods that prevent performance degradation. By resolving spatial representation dependence and optimizing for covariate shift, the study offers practical design principles essential for deploying reliable confidence estimates in medical imaging and other high-stakes computer vision tasks.
Rethinking Post-Hoc Calibration in Semantic Segmentation
arXiv:2607.01902v2 Announce Type: replace-cross Abstract: Reliable confidence estimates are essential in semantic segmentation, yet modern models often remain miscalibrated. We investigate two overlooked issues in post-hoc calibration. First, adding a constant to all logits leaves softmax probabilities unchanged, but several standard calibrators depend on this arbitrary offset. In segmentation, this offset can vary across pixels or voxels, introducing spatially varying representation dependence. We characterize translation-invariant (TI) calibrators and construct TI counterparts of shift-sensitive methods. Second, calibrating with cross-entropy can degrade segmentation quality due to mismatched training and calibration objectives and limited calibration data. We investigate decision-preserving calibration under argmax- and order-preservation constraints. Since these constraints restrict affine softmax calibrators to temperature scaling, we introduce more expressive class-conditional affine calibrators that preserve decisions. Across natural-image and medical segmentation benchmarks, including corruption-based covariate shift, TI variants generally improve calibration, while decision-preserving variants prevent segmentation degradation by construction and retain strong calibration performance. Our findings provide practical design principles for post-hoc calibration in semantic segmentation.
Authors:Tristan Kirscher (ICube, Institut Strauss), Kim-Celine Kahl (DKFZ), Balint Kovacs (DKFZ), Maximilian R. Rokuss (DKFZ), Klaus Maier-Hein (DKFZ), Xavier Coubez (ICube, Institut Strauss), Philippe Meyer (ICube, Institut Strauss), Sylvain Faisan (ICube)
Abstract:Reliable confidence estimates are essential in semantic segmentation, yet modern models often remain miscalibrated. We investigate two overlooked issues in post-hoc calibration. First, adding a constant to all logits leaves softmax probabilities unchanged, but several standard calibrators depend on this arbitrary offset. In segmentation, this offset can vary across pixels or voxels, introducing spatially varying representation dependence. We characterize translation-invariant (TI) calibrators and construct TI counterparts of shift-sensitive methods. Second, calibrating with cross-entropy can degrade segmentation quality due to mismatched training and calibration objectives and limited calibration data. We investigate decision-preserving calibration under argmax- and order-preservation constraints. Since these constraints restrict affine softmax calibrators to temperature scaling, we introduce more expressive class-conditional affine calibrators that preserve decisions. Across natural-image and medical segmentation benchmarks, including corruption-based covariate shift, TI variants generally improve calibration, while decision-preserving variants prevent segmentation degradation by construction and retain strong calibration performance. Our findings provide practical design principles for post-hoc calibration in semantic segmentation.
Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
arXiv:2607.01902 [cs.CV]
(or arXiv:2607.01902v2 [cs.CV] for this version)
Full text under the publisher's open licence · 205 words
Related coverage
Intelligence is Free, Now What? <br> Data Systems for, of, and by Agents
BAIR Blog · Jul 7, 2026
A Free Knob: Decoupling Calibration and Predictive Skill in Threshold-Based Evaluation
arXiv · Sep 30, 2026
Post-Operative Glioma Segmentation via Loss Stabilization, Normalization and Subspace Attention
arXiv · Jul 28, 2026
PTQ4SNN: Membrane-Aware Post-Training Quantization for Spiking Neural Networks
arXiv · Sep 24, 2026