Search papers, labs, and topics across Lattice.
This paper investigates the limitations of post-hoc calibration methods in semantic segmentation, particularly focusing on the impact of arbitrary logit offsets and the mismatch between calibration objectives and task-specific metrics. The authors introduce translation-invariant (TI) calibrators that maintain consistent outputs regardless of logit shifts and develop class-conditional affine calibrators that preserve decision integrity while enhancing expressivity. Their findings demonstrate that TI calibrators improve calibration metrics across various benchmarks, while decision-preserving approaches prevent degradation in segmentation quality, offering practical guidelines for implementing effective calibration in safety-critical applications.
Calibration methods that ignore logit shifts can lead to misleading confidence estimates, but translation-invariant calibrators provide a solution that enhances reliability in semantic segmentation.
Reliable confidence estimates are essential in semantic segmentation, especially in safety-critical settings where overconfident errors can mislead downstream decisions. Yet modern segmentation models often remain miscalibrated. Post-hoc calibration offers a practical way to correct confidence estimates without retraining the segmentation model, but its use in dense prediction raises structural issues that are often overlooked. We study two such issues. First, adding a constant to all logits leaves the softmax probabilities unchanged, but several standard calibrators can still depend on this arbitrary offset. As a result, two logit representations encoding the same predictive distribution may yield different calibrated probabilities. We define translation-invariant (TI) calibrators as those whose outputs are unchanged under such shifts, characterize which common calibrators satisfy this property, and construct TI counterparts of shift-sensitive calibrators to isolate the effect of removing representation dependence. Second, post-hoc calibration is typically fitted by minimizing a likelihood-based objective, whereas segmentation models are trained with task-specific metrics such as Dice. This mismatch can cause calibration to alter class orderings and degrade the deployed segmentation map. We study decision-preserving calibration under argmax- and order-preservation constraints. Since enforcing these constraints collapses affine softmax calibrators to temperature scaling, we introduce class-conditional affine calibrators that can be made argmax- or order-preserving while retaining greater expressivity, allowing us to quantify the calibration-segmentation trade-off induced by decision preservation. Across natural-image and medical segmentation benchmarks, and under corruption-based covariate shift, matched comparisons show that TI variants generally improve calibration metrics, while decision-preserving variants prevent segmentation degradation and retain strong calibration performance. These results provide practical design principles for well-defined post-hoc calibration pipelines in semantic segmentation.
