{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,13]],"date-time":"2026-04-13T07:34:23Z","timestamp":1776065663341,"version":"3.50.1"},"reference-count":44,"publisher":"IOP Publishing","issue":"1","license":[{"start":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T00:00:00Z","timestamp":1738108800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T00:00:00Z","timestamp":1738108800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"name":"\u00c5ke Wiberg Stiftelse","award":["M22-0088"],"award-info":[{"award-number":["M22-0088"]}]},{"name":"ITEA\/VINNOVA project ASSIST","award":["2021-01954"],"award-info":[{"award-number":["2021-01954"]}]},{"name":"LiU cancer"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2025,3,31]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Diffusion models were initially developed for text-to-image generation and are now being utilized to generate high quality synthetic images. Preceded by generative adversarial networks (GANs), diffusion models have shown impressive results using various evaluation metrics. However, commonly used metrics such as Frechet inception distance and inception score are not suitable for determining whether diffusion models are simply reproducing the training images. Here we train StyleGAN and a diffusion model, using BRATS20, BRATS21 and a chest x-ray (CXR) pneumonia dataset, to synthesize brain MRI and CXR images, and measure the correlation between the synthetic images and all training images. Our results show that diffusion models are more likely to memorize the training images, compared to StyleGAN, especially for small datasets and when using 2D slices from 3D volumes. Researchers should be careful when using diffusion models (and to some extent GANs) for medical imaging, if the final goal is to share the synthetic images.<\/jats:p>","DOI":"10.1088\/2632-2153\/ad9a3a","type":"journal-article","created":{"date-parts":[[2024,12,3]],"date-time":"2024-12-03T23:01:35Z","timestamp":1733266895000},"page":"015022","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":12,"title":["Beware of diffusion models for synthesizing medical images\u2014a comparison with GANs in terms of memorizing brain MRI and chest x-ray images"],"prefix":"10.1088","volume":"6","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-3248-5132","authenticated-orcid":true,"given":"Muhammad","family":"Usman Akbar","sequence":"first","affiliation":[]},{"given":"Wuhao","family":"Wang","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7061-7995","authenticated-orcid":true,"given":"Anders","family":"Eklund","sequence":"additional","affiliation":[]}],"member":"266","published-online":{"date-parts":[[2025,1,29]]},"reference":[{"key":"mlstad9a3abib1","first-page":"pp 308","article-title":"Deep learning with differential privacy","author":"Abadi","year":"2016"},{"key":"mlstad9a3abib2","doi-asserted-by":"publisher","DOI":"10.1038\/s41597-024-03073-x","article-title":"Brain tumor segmentation using synthetic MR images-A comparison of GANs and diffusion models","volume":"259","author":"Akbar","year":"2024","journal-title":"Sci. Data"},{"key":"mlstad9a3abib3","first-page":"pp 32","article-title":"Spot the fake lungs: generating synthetic medical images using neural diffusion models","author":"Ali","year":"2023"},{"key":"mlstad9a3abib4","doi-asserted-by":"publisher","DOI":"10.1038\/sdata.2017.117","article-title":"Advancing the cancer genome atlas glioma MRI collections with expert segmentation labels and radiomic features","volume":"4","author":"Bakas","year":"2017","journal-title":"Sci. Data"},{"key":"mlstad9a3abib5","article-title":"Identifying the best machine learning algorithms for brain tumor segmentation, progression assessment, and overall survival prediction in the BRATS challenge","author":"Bakas","year":"2018"},{"key":"mlstad9a3abib6","doi-asserted-by":"publisher","DOI":"10.7937\/K9\/TCIA.2017.KLXWJJ1Q","article-title":"Segmentation labels and radiomic features for the pre-operative scans of the TCGA-GBM collection","author":"Bakas","year":"2017","journal-title":"The Cancer Imaging Archive"},{"key":"mlstad9a3abib7","doi-asserted-by":"publisher","DOI":"10.7937\/K9\/TCIA.2017.GJQ7R0EF","article-title":"Segmentation labels and radiomic features for the pre-operative scans of the TCGA-LGG collection","author":"Bakas","year":"2017","journal-title":"The Cancer Imaging Archive"},{"key":"mlstad9a3abib8","doi-asserted-by":"publisher","DOI":"10.1161\/CIRCOUTCOMES.118.005122","article-title":"Privacy-preserving generative deep neural networks support clinical data sharing","volume":"12","author":"Beaulieu-Jones","year":"2019","journal-title":"Circ: Cardiovascular Qual. Outcomes"},{"key":"mlstad9a3abib9","doi-asserted-by":"publisher","first-page":"41","DOI":"10.1016\/j.cviu.2018.10.009","article-title":"Pros and cons of GAN evaluation measures","volume":"179","author":"Borji","year":"2019","journal-title":"Comput. Vis. Image Underst."},{"key":"mlstad9a3abib10","doi-asserted-by":"publisher","DOI":"10.1016\/j.cviu.2021.103329","article-title":"Pros and cons of GAN evaluation measures: new developments","volume":"215","author":"Borji","year":"2022","journal-title":"Comput. Vis. Image Underst."},{"key":"mlstad9a3abib11","article-title":"Large scale GAN training for high fidelity natural image synthesis","author":"Brock","year":"2018"},{"key":"mlstad9a3abib12","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2021.102125","article-title":"Deep learning for chest x-ray analysis: a survey","volume":"72","author":"\u00c7all\u0131","year":"2021","journal-title":"Med. Image Anal."},{"key":"mlstad9a3abib13","article-title":"Extracting training data from diffusion models","author":"Carlini","year":"2023"},{"key":"mlstad9a3abib14","doi-asserted-by":"publisher","first-page":"e29","DOI":"10.1002\/brx2.29","article-title":"Understanding the brain with attention: a survey of transformers in brain sciences","volume":"1","author":"Chen","year":"2023","journal-title":"Brain-X"},{"key":"mlstad9a3abib15","article-title":"Investigating data memorization in 3D latent diffusion models for medical image synthesis","author":"Dar","year":"2023"},{"key":"mlstad9a3abib16","first-page":"pp 8780","article-title":"Diffusion models beat GANs on image synthesis","volume":"vol 34","author":"Dhariwal","year":"2021"},{"key":"mlstad9a3abib17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2104.11797","article-title":"Ensembles of GANs for synthetic training data generation","author":"Eilertsen","year":"2021"},{"key":"mlstad9a3abib18","doi-asserted-by":"publisher","first-page":"ooab012","DOI":"10.1093\/jamiaopen\/ooab012","article-title":"Evaluating the utility of synthetic COVID-19 case data","volume":"4","author":"El Emam","year":"2021","journal-title":"JAMIA Open"},{"key":"mlstad9a3abib19","article-title":"Generative adversarial networks","author":"Goodfellow","year":"2014"},{"key":"mlstad9a3abib20","article-title":"GANs trained by a two time-scale update rule converge to a local Nash equilibrium","volume":"vol 30","author":"Heusel","year":"2017"},{"key":"mlstad9a3abib21","first-page":"pp 6840","article-title":"Denoising diffusion probabilistic models","volume":"vol 33","author":"Ho","year":"2020"},{"key":"mlstad9a3abib22","first-page":"pp 4401","article-title":"A style-based generator architecture for generative adversarial networks","author":"Karras","year":"2019"},{"key":"mlstad9a3abib23","article-title":"Alias-free generative adversarial network","author":"Karras","year":"2021"},{"key":"mlstad9a3abib24","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2023.102846","article-title":"Diffusion models in medical imaging: a comprehensive survey","volume":"textbf88","author":"Kazerouni","year":"2023","journal-title":"Med. Image Anal."},{"key":"mlstad9a3abib25","doi-asserted-by":"publisher","first-page":"1122","DOI":"10.1016\/j.cell.2018.02.010","article-title":"Identifying medical diagnoses and treatable diseases by image-based deep learning","volume":"172","author":"Kermany","year":"2018","journal-title":"Cell"},{"key":"mlstad9a3abib26","doi-asserted-by":"publisher","first-page":"7303","DOI":"10.1038\/s41598-023-34341-2","article-title":"Denoising diffusion probabilistic models for 3D medical image generation","volume":"13","author":"Khader","year":"2023","journal-title":"Sci. Rep."},{"key":"mlstad9a3abib27","article-title":"Does an ensemble of GANs lead to better performance when training segmentation networks with synthetic images?","author":"Larsson","year":"2022"},{"key":"mlstad9a3abib28","doi-asserted-by":"publisher","first-page":"1993","DOI":"10.1109\/TMI.2014.2377694","article-title":"The multimodal brain tumor image segmentation benchmark (BRATS)","volume":"34","author":"Menze","year":"2014","journal-title":"IEEE Trans. Med. Imaging"},{"key":"mlstad9a3abib29","first-page":"pp 2000","article-title":"A morphology focused diffusion probabilistic model for synthesis of histopathology images","author":"Moghadam","year":"2023"},{"key":"mlstad9a3abib30","article-title":"Generation of anonymous chest radiographs using latent diffusion models for training thoracic abnormality classification systems","author":"Packh\u00e4user","year":"2022"},{"key":"mlstad9a3abib31","doi-asserted-by":"publisher","DOI":"10.1088\/1361-6560\/acca5c","article-title":"2D medical image synthesis using transformer-based denoising diffusion probabilistic model","volume":"68","author":"Pan","year":"2023","journal-title":"Phys. Med. Biol."},{"key":"mlstad9a3abib32","doi-asserted-by":"publisher","first-page":"2518","DOI":"10.24294\/mipt.v6i1.2518","article-title":"2D brain MRI image synthesis based on lightweight denoising diffusion probabilistic model","volume":"6","author":"Peng","year":"2023","journal-title":"Med. Imaging Process Technol."},{"key":"mlstad9a3abib33","first-page":"pp 117","article-title":"Brain imaging generation with latent diffusion models","author":"Pinaya","year":"2022"},{"key":"mlstad9a3abib34","doi-asserted-by":"publisher","DOI":"10.1016\/j.isci.2022.105331","article-title":"Synthetic data as an enabler for machine learning applications in medicine","volume":"25","author":"Rajotte","year":"2022","journal-title":"Iscience"},{"key":"mlstad9a3abib35","article-title":"Hierarchical text-conditional image generation with clip latents","author":"Ramesh","year":"2022"},{"key":"mlstad9a3abib36","first-page":"pp 8821","article-title":"Zero-shot text-to-image generation","author":"Ramesh","year":"2021"},{"key":"mlstad9a3abib37","doi-asserted-by":"publisher","DOI":"10.2196\/18910","article-title":"Reliability of supervised machine learning using synthetic data in health care: model to preserve privacy for data sharing","volume":"8","author":"Rankin","year":"2020","journal-title":"JMIR Med. Inform."},{"key":"mlstad9a3abib38","first-page":"pp 10684","article-title":"High-resolution image synthesis with latent diffusion models","author":"Rombach","year":"2022"},{"key":"mlstad9a3abib39","first-page":"pp 36479","article-title":"Photorealistic text-to-image diffusion models with deep language understanding","volume":"vol 35","author":"Saharia","year":"2022"},{"key":"mlstad9a3abib40","article-title":"Improved techniques for training GANs","volume":"vol 29","author":"Salimans","year":"2016"},{"key":"mlstad9a3abib41","first-page":"pp 6048","article-title":"Diffusion art or digital forgery? Investigating data replication in diffusion models","author":"Somepalli","year":"2023"},{"key":"mlstad9a3abib42","article-title":"On the interpretability of regularisation for neural networks through model gradient similarity","author":"Szolnoky","year":"2022"},{"key":"mlstad9a3abib43","first-page":"pp 335","article-title":"Diffusion models for realistic CT image generation","author":"Txurio","year":"2023"},{"key":"mlstad9a3abib44","doi-asserted-by":"publisher","DOI":"10.1016\/j.media.2019.101552","article-title":"Generative adversarial network in medical imaging: a review","volume":"58","author":"Yi","year":"2019","journal-title":"Medical Image Analysis"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,29]],"date-time":"2025-01-29T11:27:40Z","timestamp":1738150060000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ad9a3a"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,1,29]]},"references-count":44,"journal-issue":{"issue":"1","published-online":{"date-parts":[[2025,1,29]]},"published-print":{"date-parts":[[2025,3,31]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/ad9a3a","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,1,29]]},"assertion":[{"value":"Beware of diffusion models for synthesizing medical images\u2014a comparison with GANs in terms of memorizing brain MRI and chest x-ray images","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2025 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2024-01-11","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2024-12-03","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-01-29","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}