SACP: Spatially-Adaptive Conformal Prediction in Uncertainty Quantification of Medical Image Segmentation

Jacqueline Isabel Bereska, Hamed Karimi, Reza Samavi
Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, PMLR 301:127-152, 2026.

Abstract

While Conformal Prediction provides statistical coverage guarantees, existing non-conformity measures fail to account for spatially varying importance of predictive uncertainty in medical image segmentation. In this paper, we incorporate spatial context near critical interfaces such as a vessel or critical organ in medical image segmentation. Our framework consists of three key components: (1) a base non-conformity score derived from segmentation model probabilities, (2) employing class-conditional calibration followed by a validation mechanism equipped with a distance-weighted scoring function that exponentially decays with distance from key interfaces, and (3) a prediction set construction method that preserves coverage guarantees while providing targeted uncertainty quantification in critical regions.While our approach is generalizable to different scenarios, for validation purposes, we employ tumor segmentation in pancreatic adenocarcinoma imaging from multiple medical centers. Results demonstrate that our method achieves the desired coverage levels while generating prediction sets that adaptively expand near critical interfaces.

Cite this Paper

BibTeX
@InProceedings{pmlr-v301-bereska26a, title = {SACP: Spatially-Adaptive Conformal Prediction in Uncertainty Quantification of Medical Image Segmentation}, author = {Bereska, Jacqueline Isabel and Karimi, Hamed and Samavi, Reza}, booktitle = {Proceedings of The 8th International Conference on Medical Imaging with Deep Learning}, pages = {127--152}, year = {2026}, editor = {Tasdizen, Tolga and Elhabian, Shireen and Summers, Ronald and Chen, Chen and Koch, Lisa and Zhuang, Yan}, volume = {301}, series = {Proceedings of Machine Learning Research}, month = {09--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v301/main/assets/bereska26a/bereska26a.pdf}, url = {https://proceedings.mlr.press/v301/bereska26a.html}, abstract = {While Conformal Prediction provides statistical coverage guarantees, existing non-conformity measures fail to account for spatially varying importance of predictive uncertainty in medical image segmentation. In this paper, we incorporate spatial context near critical interfaces such as a vessel or critical organ in medical image segmentation. Our framework consists of three key components: (1) a base non-conformity score derived from segmentation model probabilities, (2) employing class-conditional calibration followed by a validation mechanism equipped with a distance-weighted scoring function that exponentially decays with distance from key interfaces, and (3) a prediction set construction method that preserves coverage guarantees while providing targeted uncertainty quantification in critical regions.While our approach is generalizable to different scenarios, for validation purposes, we employ tumor segmentation in pancreatic adenocarcinoma imaging from multiple medical centers. Results demonstrate that our method achieves the desired coverage levels while generating prediction sets that adaptively expand near critical interfaces.} }
Endnote
%0 Conference Paper %T SACP: Spatially-Adaptive Conformal Prediction in Uncertainty Quantification of Medical Image Segmentation %A Jacqueline Isabel Bereska %A Hamed Karimi %A Reza Samavi %B Proceedings of The 8th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Tolga Tasdizen %E Shireen Elhabian %E Ronald Summers %E Chen Chen %E Lisa Koch %E Yan Zhuang %F pmlr-v301-bereska26a %I PMLR %P 127--152 %U https://proceedings.mlr.press/v301/bereska26a.html %V 301 %X While Conformal Prediction provides statistical coverage guarantees, existing non-conformity measures fail to account for spatially varying importance of predictive uncertainty in medical image segmentation. In this paper, we incorporate spatial context near critical interfaces such as a vessel or critical organ in medical image segmentation. Our framework consists of three key components: (1) a base non-conformity score derived from segmentation model probabilities, (2) employing class-conditional calibration followed by a validation mechanism equipped with a distance-weighted scoring function that exponentially decays with distance from key interfaces, and (3) a prediction set construction method that preserves coverage guarantees while providing targeted uncertainty quantification in critical regions.While our approach is generalizable to different scenarios, for validation purposes, we employ tumor segmentation in pancreatic adenocarcinoma imaging from multiple medical centers. Results demonstrate that our method achieves the desired coverage levels while generating prediction sets that adaptively expand near critical interfaces.
APA
Bereska, J.I., Karimi, H. & Samavi, R.. (2026). SACP: Spatially-Adaptive Conformal Prediction in Uncertainty Quantification of Medical Image Segmentation. Proceedings of The 8th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 301:127-152 Available from https://proceedings.mlr.press/v301/bereska26a.html.

Related Material