# Post Training Uncertainty Calibration of Deep Networks For Medical Image Segmentation

arXiv: https://arxiv.org/abs/2010.14290 · HTML: https://mlsanity.com/paper/2010.14290/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 27 Oct 2020.

## Built on by (most cited first, 19 of 19)

- [Trustworthy clinical AI solutions: a unified review of uncertainty quantification in deep learning models for medical image analysis](https://mlsanity.com/paper/2210.03736/) (2022)
- [A review of uncertainty quantification in medical image analysis: probabilistic and non-probabilistic methods](https://mlsanity.com/paper/2310.06873/) (2023)
- [Multimodal Distillation for Egocentric Action Recognition](https://mlsanity.com/paper/2307.07483/) (2023)
- [Conformal Semantic Image Segmentation: Post-hoc Quantification of Predictive Uncertainty](https://mlsanity.com/paper/2405.05145/) (2024)
- [Calibrating the Dice loss to handle neural network overconfidence for biomedical image segmentation](https://mlsanity.com/paper/2111.00528/) (2021)
- [Lymphoma segmentation from 3D PET-CT images using a deep evidential network](https://mlsanity.com/paper/2201.13078/) (2022)
- [Jaccard Metric Losses: Optimizing the Jaccard Index with Soft Labels](https://mlsanity.com/paper/2302.05666/) (2023)
- [On the relationship between calibrated predictors and unbiased volume estimation](https://mlsanity.com/paper/2112.12560/) (2021)
- [On the Optimal Combination of Cross-Entropy and Soft Dice Losses for Lesion Segmentation with Out-of-Distribution Robustness](https://mlsanity.com/paper/2209.06078/) (2022)
- [On the Overconfidence Problem in Semantic 3D Mapping](https://mlsanity.com/paper/2311.10018/) (2023)
- [On Image Segmentation With Noisy Labels: Characterization and Volume Properties of the Optimal Solutions to Accuracy and Dice](https://mlsanity.com/paper/2206.06484/) (2022)
- [Dice Semimetric Losses: Optimizing the Dice Score with Soft Labels](https://mlsanity.com/paper/2303.16296/) (2023)
- [Uncertainty categories in medical image segmentation: a study of source-related diversity](https://mlsanity.com/paper/2203.00238/) (2022)
- [Deep evidential fusion with uncertainty quantification and contextual discounting for multimodal medical image segmentation](https://mlsanity.com/paper/2309.05919/) (2023)
- [Noisy Image Segmentation With Soft-Dice](https://mlsanity.com/paper/2304.00801/) (2023)
- [Uncertainty Quantification in Machine Learning Based Segmentation: A Post-Hoc Approach for Left Ventricle Volume Estimation in MRI](https://mlsanity.com/paper/2312.02167/) (2023)
- [Marginal Thresholding in Noisy Image Segmentation](https://mlsanity.com/paper/2304.04116/) (2023)
- [Can ultrasound confidence maps predict sonographers' labeling variability?](https://mlsanity.com/paper/2308.09433/) (2023)
- [Beyond scalar losses: calibrating segmentation models via gradient vector field surgery](https://mlsanity.com/paper/2607.14338/) (2026)

## Abstract (arXiv)

Neural networks for automated image segmentation are typically trained to achieve maximum accuracy, while less attention has been given to the calibration of their confidence scores. However, well-calibrated confidence scores provide valuable information towards the user. We investigate several post hoc calibration methods that are straightforward to implement, some of which are novel. They are compared to Monte Carlo (MC) dropout. They are applied to neural networks trained with cross-entropy (CE) and soft Dice (SD) losses on BraTS 2018 and ISLES 2018. Surprisingly, models trained on SD loss are not necessarily less calibrated than those trained on CE loss. In all cases, at least one post hoc method improves the calibration. There is limited consistency across the results, so we can't conclude on one method being superior. In all cases, post hoc calibration is competitive with MC dropout. Although average calibration improves compared to the base model, subject-level variance of the calibration remains similar.