mlsanityData to 7 Oct 2026

Investigating Data Memorization in 3D Latent Diffusion Models for Medical Image Synthesis

arXiv 2307.01148 ↗First posted 3 Jul 2023cs.CV eess.IVcited by 19

On the map: Diffusion models → Medical imaging and segmentation

Abstract (arXiv)

Generative latent diffusion models have been established as state-of-the-art in data generation. One promising application is generation of realistic synthetic medical imaging data for open data sharing without compromising patient privacy. Despite the promise, the capacity of such models to memorize sensitive patient training data and synthesize samples showing high resemblance to training data samples is relatively unexplored. Here, we assess the memorization capacity of 3D latent diffusion models on photon-counting coronary computed tomography angiography and knee magnetic resonance imaging datasets. To detect potential memorization of training samples, we utilize self-supervised models based on contrastive learning. Our results suggest that such latent diffusion models indeed memorize training data, and there is a dire need for devising strategies to mitigate memorization.

Key facts. Cited by 19 papers in the map (data through 7 Oct 2026). Placed in Medical imaging and segmentation, one of 14 directions in Diffusion models. Builds on 15 of its 19 references that are in the map. First posted on arXiv 3 Jul 2023.

Builds on

15 of 19 references.

Closest ancestors · 6

Same direction, newest first

  1. 2022Diffusion Models for Medical Image Analysis: A Comprehensive Survey14 Nov
  2. 2022Brain Imaging Generation with Latent Diffusion Models15 Sep
  3. 2022Unsupervised Medical Image Translation with Adversarial Diffusion Models17 Jul
  4. 2022Adaptive Diffusion Priors for Accelerated MRI Reconstruction12 Jul
  5. 2022Diffusion Models for Medical Anomaly Detection8 Mar
  6. 2021Diffusion Models for Implicit Image Segmentation Ensembles6 Dec

Other references · 9

  1. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  3. 2015FaceNet: A Unified Embedding for Face Recognition and Clusteringcited by 2,833
  4. 2023Extracting Training Data from Diffusion Modelscited by 723
  5. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 418
  6. 2018Generative Adversarial Network in Medical Imaging: A Reviewcited by 257
  7. 2023Understanding and Mitigating Copying in Diffusion Modelscited by 205
  8. 2019Generating large labeled data sets for laparoscopic image processing tasks using unpaired image-to-image translationcited by 38
  9. 2019Cross-Domain Conditional Generative Adversarial Networks for Stereoscopic Hyperrealism in Surgical Trainingcited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 4; 2024: 5; 2025: 9; 2026: 1.

19 papers.

Sort:

By area:

  1. 2023Generalization in diffusion models arises from geometry-adaptive harmonic representationscited by 179
  2. 2025On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticitycited by 48
  3. 2025Generalization through variance: how noise shapes inductive biases in diffusion modelscited by 23
  4. 2023On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlookcited by 15
  5. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  6. 2023A Survey of Emerging Applications of Diffusion Probabilistic Models in MRIcited by 9
  7. 2024Unconditional Latent Diffusion Models Memorize Patient Imaging Data: Implications for Openly Sharing Synthetic Datacited by 9
  8. 2024Synthetic Data in Radiological Imaging: Current State and Future Outlookcited by 7
  9. 2025Navigating the Latent Space Dynamics of Neural Modelscited by 7
  10. 2025CheXGenBench: A Unified Benchmark For Fidelity, Privacy and Utility of Synthetic Chest Radiographscited by 5
  11. 2025Diffusion models under low-noise regimecited by 5
  12. 2025LCMem: A Universal Model for Robust Image Memorization Detectioncited by 4
7 more
  1. 2024MemControl: Mitigating Memorization in Diffusion Models via Automated Parameter Selectioncited by 3
  2. 2024Latent Pollution Model: The Hidden Carbon Footprint in 3D Image Synthesiscited by 2
  3. 2025The Devil is in the Prompts: De-Identification Traces Enhance Memorization Risks in Synthetic Chest X-Ray Generationcited by 2
  4. 2025A Novel Metric for Detecting Memorization in Generative Models for Brain MRI Synthesiscited by 2
  5. 2025On the Collapse Errors Induced by the Deterministic Sampler for Diffusion Modelscited by 1
  6. 2023Trustworthy Large Models in Vision: A Surveycited by 0
  7. 2026Auditing Patient Privacy in Medical Generative Models: Scalable Memorization Detection with DeepSSIM++cited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026The Learnability Gap in Medical Latent Diffusion
  2. 2022Medical Diffusion: Denoising Diffusion Probabilistic Models for 3D Medical Image Generation
  3. 2025Lesion-Aware Post-Training of Latent Diffusion Models for Synthesizing Diffusion MRI from CT Perfusion
  4. 2024Latent Drifting in Diffusion Models for Counterfactual Medical Image Synthesis
  5. 2025Computationally Efficient Diffusion Models in Medical Imaging: A Comprehensive Review
  6. 2025Diffusion Models for conditional MRI generation
  7. 2024Diffusion-Based Approaches in Medical Image Generation and Analysis
  8. 2026What Makes a Good Medical Image Tokenizer? Rethinking Reconstruction and Generation in Medical Image Tokenization