PuSH - Publication Server of Helmholtz Zentrum München

Konz, N.* ; Osuala, R. ; Verma, P.* ; Chen, Y.* ; Gu, H.* ; Dong, H.* ; Chen, Y.* ; Marshall, A.* ; Garrucho, L.* ; Kushibar, K.* ; Lang, D. ; Kim, G.S.* ; Grimm, L.J.* ; Lewin, J.M.* ; Duncan, J.S.* ; Schnabel, J.A. ; Diaz, O.* ; Lekadir, K.* ; Mazurowski, M.A.*

Fréchet radiomic distance (FRD): A versatile metric for comparing medical imaging datasets.

Med. Image Anal. 110:103943 (2026)
Postprint DOI PMC
Open Access Green
Determining whether two sets of images belong to the same or different distributions or domains is a crucial task in modern medical image analysis and deep learning; for example, to evaluate the output quality of image generative models. Currently, metrics used for this task either rely on the (potentially biased) choice of some downstream task, such as segmentation, or adopt task-independent perceptual metrics (e.g., Fréchet Inception Distance/FID) from natural imaging, which we show insufficiently capture anatomical features. To this end, we introduce a new perceptual metric tailored for medical images, FRD (Fréchet Radiomic Distance), which utilizes standardized, clinically meaningful, and interpretable image features. We show that FRD is superior to other image distribution metrics for a range of medical imaging applications, including out-of-domain (OOD) detection, the evaluation of image-to-image translation (by correlating more with downstream task performance as well as anatomical consistency and realism), and the evaluation of unconditional image generation. Moreover, FRD offers additional benefits such as stability and computational efficiency at low sample sizes, sensitivity to image corruptions and adversarial attacks, feature interpretability, and correlation with radiologist-perceived image quality. Additionally, we address key gaps in the literature by presenting an extensive framework for the multifaceted evaluation of image similarity metrics in medical imaging-including the first large-scale comparative study of generative models for medical image translation-and release an accessible codebase to facilitate future research. Our results are supported by thorough experiments spanning a variety of datasets, modalities, and downstream tasks, highlighting the broad potential of FRD for medical image analysis.
Altmetric
Additional Metrics?
[➜Log in]
Edit extra informations Login
Publication type Article: Journal article
Document type Scientific Article
Keywords Evaluation ; Generative Models ; Image Similarity ; Image-to-image Translation ; Metrics ; Ood Detection; Cancer; Images
ISSN (print) / ISBN 1361-8415
e-ISSN 1361-8415
Journal Medical Image Analysis
Quellenangaben Volume: 110, Issue: , Pages: , Article Number: 103943 Supplement: ,
Publisher Elsevier
Publishing Place Radarweg 29, 1043 Nx Amsterdam, Netherlands
Reviewing status Peer reviewed
Institute(s) Institute for Machine Learning in Biomed Imaging (IML)
Grants HELMHOLTZ IMAGING, a platform of the Helmholtz Information & Data Science Incubator
Helmholtz Information and Data Science Academy (HIDA)
AIMED from the Ministry of Science, Innovation and Universities of Spain
Project FUTURE-ES
European Union's Horizon Europe and Horizon 2020 research and innovation programme
National Institute Of Biomedical Imaging And Bioengineering of the National Institutes of Health