Conformal Risk Control for Semantic Uncertainty Quantification in Computed Tomography

Jacopo Teneggi, J Webster Stayman, Jeremias Sulam
2/28/2025
cs.CVstat.ML

No official code found. Best implementation: Sulam-Group/semantic_uq (1 star, MIT, last commit 2025-09-03). 1 more below.

Abstract

Uncertainty quantification is necessary for developers, physicians, and regulatory agencies to build trust in machine learning predictors and improve patient care. Beyond measuring uncertainty, it is crucial to express it in clinically meaningful terms that provide actionable insights. This work introduces a conformal risk control (CRC) procedure for organ-dependent uncertainty estimation, ensuring high-probability coverage of the ground-truth image. We first present a high-dimensional CRC procedure that leverages recent ideas of length minimization. We make this procedure semantically adaptive to each patient's anatomy and positioning of organs. Our method, sem-CRC, provides tighter uncertainty intervals with valid coverage on real-world computed tomography (CT) data while communicating uncertainty with clinically relevant features.

View on arXivView PDF

Code Implementations(2)

Sulam-Group/semantic_uqCommunity73%
MIT

Conformal Risk Control for Semantic Uncertainty Quantification in Computed Tomography

10Shell, PythonFeb 28, 20251 years agoMIT
OJules/Conformal-Risk-Control-Medical-AICommunity69%
MIT

Master's thesis: Conformal Risk Control for Semantic Uncertainty Quantification in Computed Tomography

00Python, Jupyter NotebookSep 24, 20251 years agoMIT
computed-tomographyhealthcare-aimaster-thesismedical-aimedical-imaging+2 more

Similar Papers

Position Paper: Integrating Explainability and Uncertainty Estimation in Medical AI

83% match2025

Statistical Management of the False Discovery Rate in Medical Instance Segmentation Based on Conformal Risk Control

83% match2025

Interpretability of Uncertainty: Exploring Cortical Lesion Segmentation in Multiple Sclerosis

83% match1 repo32024

Enhancing Neuro-Oncology Through Self-Assessing Deep Learning Models for Brain Tumor Unified Model for MRI Segmentation

82% match2025

Explaining Uncertainty in Multiple Sclerosis Lesion Segmentation Beyond Prediction Errors

82% match2025

Rate-In: Information-Driven Adaptive Dropout Rates for Improved Inference-Time Uncertainty Estimation

82% match1 repo82026

Cite this paper

@article{teneggi2025conformal,
  title  = {Conformal Risk Control for Semantic Uncertainty Quantification in Computed Tomography},
  author = {Jacopo Teneggi and J Webster Stayman and Jeremias Sulam},
  year   = {2025},
  eprint = {2503.00136},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2503.00136v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2503.00136.svg)](https://codeofpaper.com/paper/2503.00136)

Discussion