Uncertainty as a Proxy for Semantic Correctness in Diffusion-Based Medical Image Synthesis

Yuxuan Ou; Konstantinos Kamnitsas; OxAAA Study; AICT Consortium; Regent Lee; Vicente Grau

cs.CVOct 2, 2026

Uncertainty as a Proxy for Semantic Correctness in Diffusion-Based Medical Image Synthesis

Authors: Yuxuan Ou, Konstantinos Kamnitsas, OxAAA Study, AICT Consortium, Regent Lee, Vicente Grau

Organizations: Department of Engineering Science, University of Oxford, Oxford, United Kingdom · Nuffield Department of Surgical Sciences, University of Oxford, Oxford, United Kingdom

Abstract

Diffusion models can synthesise contrast-enhanced CT (CECT) from non-contrast CT (NCCT), avoiding contrast administration and its environmental and patient-access costs. However, visually realistic images are not necessarily anatomically correct, and the pixel-intensity and feature-space similarity metrics used to assess generation quality do not directly measure anatomical correctness. In this work, we investigate whether uncertainty can serve as a proxy for semantic correctness in diffusion-based medical image synthesis. We study NCCT-to-CECT synthesis using AortaDiff, a multitask diffusion framework that jointly generates CECT images and lumen segmentations. The segmentation output provides an explicit representation of the generated vascular anatomy, enabling segmentation-derived errors to be used as a quantitative measure of generation correctness. Six methods spanning weight (Ensemble, HyperDiff, BayesDiff), architecture-perturbation (MCDropout), generative-stochasticity (RDS) and input-perturbation (TTA) uncertainty are compared at the pixel, region and image levels, and for detection of clinically relevant out-of-distribution (OOD) cases. Uncertainty proves informative at all three spatial scales, remains informative on an external multi-centre dataset under distribution shift, and supports OOD detection. MCDropout stands out among the six: it ranks among the leading methods at every scale, generalizes well on the external dataset, and can be enabled at inference on any model already trained with dropout, so reliable uncertainty comes at no extra training cost. Uncertainty reliably flags severe failures but discriminates poorly among already high-quality images. These findings support uncertainty as a practical and computationally economical signal for quality filtering, reliability assessment and OOD detection in NCCT-to CECT synthesis.

Figures & tables

Fig. 1: Overview of the proposed evaluation framework. AortaDiff jointly generates CECT and lumen segmentation from NCCT. Uncertainty is evaluated against segmentation-derived semantic error at pixel, region, and image levels, with additional external OOD evaluation.

Source	Method	Mechanism and configuration	Passes	Time (s)	Training overhead
Input perturbation	TTA	400 augmented views (bounding-box expansion, rotation, isotropic scaling), inverse-warped to original coordinates	400	42	None
Generative stochasticity	RDS	Deterministic DDIM ( $\eta=0$ ) from 400 initial noise realizations	400	42	None
Architecture perturbation	MCDropout	Dropout $p=0.5$ in U-Net residual blocks; 20 seeds $\times$ 20 trajectories	400	44	None
Weight	Ensemble	4 independently trained models $\times$ 100 samples	400	44	4 models
	HyperDiff	Bayesian hypernetwork generating denoiser weights; 20 weight sets $\times$ 20 samples	400	55	Hypernetwork
	BayesDiff	Last-layer Laplace approximation (diagonal Hessian) on the final convolutional layer	1	0.22	Laplace fitting

TABLE I: Evaluated uncertainty estimation methods. All use DDIM sampling with 10 steps; generation is the mean over samples and uncertainty the pixel-wise sample variance. Inference time is per test case.

Fig. 2: Pixel-wise sparsification curves for different uncertainty estimation methods. Pixels within the aortic region are gradually removed according to their uncertainty ranking, and the retained Dice is recomputed at each step. The uncertainty-guided curve (blue) is compared against the theoretical best (green) and random removal (red).

Method	AUSE $\downarrow$	AUSC $\uparrow$	AUROC $\uparrow$
TTA	0.037	0.904	0.787 ± 0.094
Ensemble	0.063	0.878	0.689 ± 0.122
RDS	0.035	0.906	0.809 ± 0.094
MCDropout	0.028	0.911	0.830 ± 0.073
HyperDiff	0.049	0.890	0.815± 0.084
BayesDiff	0.170	0.771	0.493 ± 0.049

TABLE II: Pixel-level comparison of uncertainty estimation methods. AUSE (

\downarrow

) and AUSC (

\uparrow

) summarise the pixel-wise sparsification curve; AUROC (

\uparrow

) measures classification of correct versus incorrect pixels. Best in bold.

Fig. 3: Region-level sparsification curves. SLIC superpixels are progressively removed in order of decreasing average uncertainty and the retained Dice recomputed at each step, comparing each method’s uncertainty-guided curve (blue) against the oracle (green) and random removal (red).

Fig. 4: HyperDiff pixel-level under-flagging versus region-level uncertainty on a representative slice.

Method	AUSE $\downarrow$	AUSC $\uparrow$	Spearman $\uparrow$
MCDropout	0.023	0.901	0.65 ± 0.19
TTA	0.028	0.897	0.56 ± 0.23
RDS	0.042	0.881	0.60 ± 0.20
Ensemble	0.041	0.882	0.44 ± 0.29
HyperDiff	0.021	0.899	0.64 ± 0.16
BayesDiff	0.160	0.764	-0.24 ± 0.28

TABLE III: Region-level comparison of uncertainty estimation methods. AUSE (

\downarrow

) and AUSC (

\uparrow

) summarise the localized sparsification curve; Spearman correlation (

\uparrow

) measures rank agreement between region uncertainty and region semantic error.

Fig. 5: Image-level failure detection. A sample is labelled failed when its lumen Dice falls below a threshold; the threshold is swept from 0.6 to 1.0 and AUROC reported at each.

Fig. 6: Image-level sparsification curves. Images are progressively removed in order of decreasing aggregated uncertainty and the average Dice of the retained images recomputed at each step, compared against the oracle (green) and random removal (red).

Fig. 7: Representative OOD cases containing vascular stents that were absent from the training dataset.

Fig. 8: Image-level failure detection plots for different thresholds in AICT dataset.

Method	Mean $\uparrow$	Max $\uparrow$	P90 $\uparrow$	MWA 7 $\uparrow$
MCDropout	0.77	0.83	0.81	0.82
RDS	0.79	0.84	0.79	0.85
HyperDiff	0.75	0.78	0.75	0.81
TTA	0.80	0.85	0.81	0.83
BayesDiff	0.48	0.82	0.53	0.72
Ensemble	0.78	0.81	0.79	0.80

TABLE IV: OOD stent detection AUROC (

\uparrow

) under four aggregation strategies: mean, maximum, 90th percentile, and maximum local average with window size 7.

Medical Image Synthesis

Medical image synthesis creates, translates, reconstructs, and edits clinical images to fill missing data and support analysis.

See all 310 papers

Sep 15, 2026

MUMINS: Metadata-conditioned Uncertainty-aware Medical Image Next-state Synthesis

Forecasting anatomical changes such as tumor growth and neurodegeneration is a challenging generative vision task. Morphological evolution is subtle relative to static anatomy, highly patient-specific, and inherently stochastic. Existing methods struggle with several issues: deterministic networks ignore biological stochasticity, while standard diffusion models require computationally prohibitive multi-pass sampling to quantify uncertainty. We propose MUMINS (Metadata-conditioned Uncertainty-aware Medical Image Next-state Synthesis), an efficient diffusion framework that jointly diffuses a baseline scan and its follow-up residual, summed to synthesize the follow-up scan, while concurrently predicting a spatial uncertainty map, in a single reverse diffusion process. Conditioned on the time interval and relevant metadata, it preserves fine-grained anatomy by dynamically re-injecting the baseline as a soft anchor at every denoising step, and a negative-log-likelihood head learns the uncertainty map to explicitly flag error-prone regions. Designed without organ-specific heuristics, the same architecture is reused across anatomies via separate, dataset-specific retraining. Extensive evaluations demonstrate that dataset-specific retraining of MUMINS matches or outperforms dedicated, domain-specific state-of-the-art methods on lung CT (PNG) and brain MRI (OASIS-3). Project page: https://aolivtous.github.io/publications/mumins.html.

Anna Oliveras, Roger Marí, Rafael Redondo +7

Eurecat, Centre Tecnol`ogic de Catalunya, Barcelona, Spain · Dept. de Matem`atiques i Inform`atica, Universitat de Barcelona, Barcelona, Spain · Barcelona Supercomputing Center (BSC), Barcelona, Spain +1

Medical Image Analysis Longitudinal Medical Image Analysis Medical Image Synthesis Conditional Diffusion Models

Ask PDF

Apr 24, 2026

VS-DDPM: Efficient Low-Cost Diffusion Model for Medical Modality Translation

Diffusion models produce high-quality synthetic data but suffer from slow inference. We propose 3D Variable-Step Denoising Diffusion Probabilistic Model (VS-DDPM) a framework engineered to maintain generative quality while accelerating inference by several factors. We tested our approach on four tasks (missing MRI, tumor removal, MRI-to-sCT, and CBCT-to-sCT) within the BraTS2025 and SynthRAD2025 challenges. Designed for high efficiency under hardware and time constrains imposed by both challenges. VS-DDPM achieved state-of-the-art (SOTA) performance in missing MRI synthesis, yielding Dice scores of 0.80, 0.83, and 0.88 for the enhancing tumor, tumor core, and whole tumor regions, respectively, alongside a structural similarity index (SSIM) of 0.95. For MRI tumor removal, the model attained a root mean squared error (RMSE) of 0.053, a peak signal-to-noise ratio (PSNR) of 26.77, and an SSIM of 0.918. While the framework demonstrated competitive performance in MRI-to-sCT and CBCT-to-sCT tasks, it did not reach SOTA benchmarks, potentially due to sensitivities in data pre and post-processing pipelines or specific loss function configurations. These results demonstrate that VS-DDPM provides a robust and tunable solution for high-fidelity 3D medical image synthesis. The code is available in https://github.com/andre-fs-ferreira/SynthRAD_by_Faking_it.

Nikoo Moradi, Gijs Luijten, Behrus Hinrichs-Puladi +4

Institute for Artificial Intelligence in Medicine (IKIM), Essen University Hospital (AöR), University of Duisburg-Essen, Essen, Germany · Institute of Computer Graphics and Vision, Graz University of Technology, Inffeldgasse 16, Graz, 8010, Austria · Center for Virtual and Extended Reality in Medicine (ZvRM), University Hospital Essen (AöR), Essen, Germany +7

Image-to-Image Translation 3D Medical Imaging Few-Step Diffusion Sampling Medical Image Synthesis 3D Diffusion Models Diffusion Model Inference Acceleration

Ask PDF

Jul 22, 2026

Trustworthy Medical Segmentation: Uncertainty-Aware U-Net Evaluation Under Clinical Image Degradation

Medical image segmentation models often report high benchmark accuracy under ideal imaging conditions, yet their failures under clinical degradation can be quiet: sensor noise, patient motion, low- resolution acquisition, and contrast variability may all alter model behavior without producing an obvious warning. We present a reproducible framework for evaluating uncertainty-aware segmentation under con- trolled clinical degradation. Our experiments use a synthetic multimodal brain tumor MRI cohort generated with a biophysical phantom simulator that follows the BraTS protocol. We train U-Net and Attention U-Net baselines for multi-class tumor sub-region segmentation and augment both models with Monte Carlo dropout to estimate per-voxel uncertainty. Across eight clinically motivated corruption types at five severity levels, we measure segmentation accuracy, calibration, failure detection, and selective prediction coverage. On clean data, Attention U-Net achieves a whole-tumor Dice of 0.990; under severe Gaussian noise, its performance falls to 0.089. Predictive uncertainty rises with degradation and tracks segmentation error (Pearson r = 0.53 under severity-3 Gaussian noise), allowing us to flag failures with an AUROC of 0.843. These results argue for uncertainty-aware inference as a practical safety layer in physician-in-the-loop radiology workflows. We release the code, trained models, and evaluation protocol to support direct reproduction.

Pranav Kaliaperumal, Manisha Kaliaperumal

Department of Computer Science University of Colorado Denver Aurora, CO, USA · Department of Biology (Pre-Medical) Creighton University Omaha, NE, USA

Image Corruption Robustness U-Net Medical Image Segmentation Segmentation Uncertainty Estimation Brain Tumor Segmentation

Ask PDF

Abstract

Medical Image Synthesis

Explore similar work

MUMINS: Metadata-conditioned Uncertainty-aware Medical Image Next-state Synthesis

VS-DDPM: Efficient Low-Cost Diffusion Model for Medical Modality Translation

Trustworthy Medical Segmentation: Uncertainty-Aware U-Net Evaluation Under Clinical Image Degradation