A Method for Detecting Distinctive Patterns of Real Patients in Generated Images
https://doi.org/10.35596/1729-7648-2025-23-1-47-53
Abstract
Generative diffusion models are a well-established method for generating high-quality images. However, there are studies that show that diffusion models are less privacy-friendly than generative models, such as generative adversarial networks and a growing family of their modifications. The discovered vulnerabilities require in-depth study of various security aspects. This is especially important for sensitive areas such as medical image analysis tasks and their practical applications. The paper describes a method for detecting image patterns presented in generated images that can potentially be identified in real CT images of patients with pulmonary tuberculosis. The method includes the following main procedures: correlation of pairs of generated and real images to pre-select pairs that involve further analysis; calculation of correlation statistics using direct and inverse Fisher transforms; performing affine image registration and calculating pairwise similarity scores; nonlinear (elastic) image registration and recalculation of similarity scores to highlight the most similar/dissimilar image areas.
About the Author
V. A. Kovalev
The United Institute of Informatics Problems of the National Academy of Sciences of Belarus
Belarus
Belarus
Kovalev Vassili Alekseevich, Сand. of Sci., Leading Researcher
220013, Minsk, Surganova St., 6
References
1. Koshino K., Werner R. A., Pomper M. G., Bundschuh R. A. Toriumi F., Higuchi T., et al. (2021) Narrative Review of Generative Adversarial Networks in Medical and Molecular Imaging. Annals of Translational Medicine. 9 (9), 821–835.
2. Chambon P., Bluethgen C., Langlotz C. P., Chaudhari A. (2022) Adapting Pretrained Vision-Language Foundational Models to Medical Imaging Domains. arXiv paper arXiv: 2210.04133. 1–17.
3. Kozlovski S., Kovalev V. (2019) Generation of Artificial Biomedical Image Datasets for Training Deep Learning Models. Pattern Recognition and Information Processing, 14th International Conference, Minsk, May 21–23, Belarusian State University of Informatics and Radioelectronics. 278–281.
4. Song Y., Ermon S. (2019) Generative Modeling by Estimating Gradients of the Data Distribution. 33rd Conference on Neural Information Processing Systems, Vancouver, Canada, Dec. 8–14. 115–119.
5. Ho J., Jain A., Abbeel P. (2020) Denoising Diffusion Probabilistic Models. Advances in Neural Information Processing Systems. 33, 6840–6851.
6. Carlini N., Hayes J., Nasr M., Jagielski M., Sehwag V., Tramer F., et al. (2023) Extracting Training Data from Diffusion Models. Proceedings of the 32nd USENIX Security Symposium, Anaheim, CA, USA, Aug. 9–11. 5253–5270.
7. Fisher R. A. (1915) The Frequency Distribution of the Values of the Correlation Coefficient in Samples of an Indefinitely Large Population. Biometrika. 10 (4), 507–521.
8. R Core Team (2022) R: A Language and Environment for Statistical Computing. Vienna, Foundation for Statistical Computing.
9. Heusel M., Ramsauer H., Unterthiner T., Nessler B., Hochreiter S. (2017) GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilibrium. NIPS’17: Proceedings of the 31st International Conference on Neural Information Processing Systems. 6629–6640.
10. Ruthotto L., Modersitzki J. (2015) Non-Linear Image Registration. Handbook of Mathematical Methods in Imaging. Springer, New York.
11. Cover T. M., Thomas J. A. (1991) Elements of Information Theory. New York, John Wiley & Sons.
12. Maes F., Loeckx D., Vandermeulen D., Suetens P. (2015) Image Registration Using Mutual Information. Handbook of Biomedical Imaging. Springer. 295–308.
Review
For citations:
Kovalev V.A. A Method for Detecting Distinctive Patterns of Real Patients in Generated Images. Doklady BGUIR. 2025;23(1):47-53. https://doi.org/10.35596/1729-7648-2025-23-1-47-53
Views: 48
Email this article 