Abstract
The lack of interpretability of Deep Learning models hinders their deployment in clinical contexts. Case-based explanations can be used to justify these models’ decisions and improve their trustworthiness. However, providing medical cases as explanations may threaten the privacy of patients. We propose a generative adversarial network to disentangle identity and medical features from images. Using this network, we can alter the identity of an image to anonymize it while preserving relevant explanatory features. As a proof of concept, we apply the proposed model to biometric and medical datasets, demonstrating its capacity to anonymize medical images while preserving explanatory evidence and a reasonable level of intelligibility. Finally, we demonstrate that the model is inherently capable of generating counterfactual explanations.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Alber, M., et al.: iNNvestigate neural networks! J. Mach. Learn. Res. 20(93), 1–8 (2019)
Bromley, J., et al.: Signature verification using a “siamese’’ time delay neural network. Int. J. Pattern Recognit. Artif. Intell. 7, 25 (1993). https://doi.org/10.1142/S0218001493000339
Chen, J., Konrad, J., Ishwar, P.: VGAN-based image representation learning for privacy-preserving facial expression recognition. In: CVPR, CV-COPS (2018)
Cho, D., Lee, J.H., Suh, I.H.: CLEANIR: controllable attribute-preserving natural identity remover. Appl. Sci. 10(3), 1120 (2020)
DeGrave, A.J., Janizek, J.D., Lee, S.: AI for radiographic COVID-19 detection selects shortcuts over signal. Nat. Mach. Intell. 3(7), 610–619 (2021)
Ghimire, S., Kashyap, S., Wu, J.T., Karargyris, A., Moradi, M.: Learning invariant feature representation to improve generalization across chest x-ray datasets. In: Machine Learning in Medical Imaging, pp. 644–653 (2020)
Gong, M., Liu, J., Li, H., Xie, Y., Tang, Z.: Disentangled representation learning for multiple attributes preserving face deidentification. IEEE Trans. Neural Netw. Learn. Syst. 1–13 (2020). https://doi.org/10.1109/TNNLS.2020.3027617
Goodfellow, I., et al.: Generative adversarial nets. In: Advances in Neural Information Processing Systems, vol. 27, pp. 2672–2680 (2014)
Gross, R., Airoldi, E., Malin, B., Sweeney, L.: Integrating utility into face de-identification. In: Danezis, G., Martin, D. (eds.) PET 2005. LNCS, vol. 3856, pp. 227–242. Springer, Heidelberg (2006). https://doi.org/10.1007/11767831_15
Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., Courville, A.: Improved training of Wasserstein GANs. In: NIPS 2017, pp. 5769–5779 (2017)
Hadsell, R., Chopra, S., LeCun, Y.: Dimensionality reduction by learning an invariant mapping. In: 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2006), vol. 2, pp. 1735–1742 (2006). https://doi.org/10.1109/CVPR.2006.100
Irvin, J., et al.: Chexpert: a large chest radiograph dataset with uncertainty labels and expert comparison (2019)
Irvin, J., et al.: Chexpert: a large chest X-ray dataset and competition (2019). https://stanfordmlgroup.github.io/competitions/chexpert/. Accessed 21 Feb 2022
Kingma, D.P., Welling, M.: Auto-encoding variational bayes. In: ICLR (2013)
Lipton, Z.C.: The mythos of model interpretability: in machine learning, the concept of interpretability is both important and slippery. Queue 16(3), 31–57 (2018)
Mahmood, U., et al.: Detecting spurious correlations with sanity tests for artificial intelligence guided radiology systems. Front. Digit. Health 3, 85 (2021)
Montavon, G., Lapuschkin, S., Binder, A., Samek, W., Müller, K.: Explaining nonlinear classification decisions with deep taylor decomposition. Pattern Recog. 65, 211–222 (2017)
Montenegro, H., Silva, W., Cardoso, J.S.: Privacy-preserving generative adversarial network for case-based explainability in medical image analysis. IEEE Access 9, 148037–148047 (2021)
Montenegro, H., Silva, W., Cardoso, J.S.: Towards privacy-preserving explanations in medical image analysis. In: IMLH Workshop, at ICML 2021 (2021)
Montenegro, H., Silva, W., Gaudio, A., Fredrikson, M., Smailagic, A., Cardoso, J.S.: Privacy-preserving case-based explanations: enabling visual interpretability by protecting privacy. IEEE Access 10, 28333–28347 (2022). https://doi.org/10.1109/ACCESS.2022.3157589
Oleszkiewicz, W., Kairouz, P., Piczak, K., Rajagopal, R., Trzciński, T.: Siamese generative adversarial privatizer for biometric data. In: Jawahar, C.V., Li, H., Mori, G., Schindler, K. (eds.) ACCV 2018. LNCS, vol. 11365, pp. 482–497. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-20873-8_31
Schölkopf, B.: Causality for machine learning. arXiv preprint arXiv:1911.10500 (2019)
Schölkopf, B., et al.: Towards causal representation learning (2021). https://doi.org/10.48550/ARXIV.2102.11107. https://arxiv.org/abs/2102.11107
Trokielewicz, M., Czajka, A., Maciejewicz, P.: Assessment of iris recognition reliability for eyes affected by ocular pathologies. In: BTAS Conference, pp. 1–6 (2015)
Trokielewicz, M., Czajka, A., Maciejewicz, P.: Biometric databases (2015). http://zbum.ia.pw.edu.pl/EN/node/46. Accessed 18 Oct 2021
Trokielewicz, M., Czajka, A., Maciejewicz, P.: Implications of ocular pathologies for iris recognition reliability. Image Vis. Comput. 58, 158–167 (2017)
Wu, Y., Yang, F., Xu, Y., Ling, H.: Privacy-protective-GAN for privacy preserving face de-identification. J. Comput. Sci. Technol. 34(1), 47–60 (2019). https://doi.org/10.1007/s11390-019-1898-8
Acknowledgements
This work was partially funded by the Project TAMI - Transparent Artificial Medical Intelligence (NORTE-01-0247-FEDER-045905) financed by ERDF - European Regional Fund through the North Portugal Regional Operational Program - NORTE 2020 and by the Portuguese Foundation for Science and Technology - FCT under the CMU - Portugal International Partnership, and also by the Portuguese Foundation for Science and Technology - FCT within PhD grant number SFRH/BD/139468/2018.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Montenegro, H., Silva, W., Cardoso, J.S. (2023). Disentangled Representation Learning for Privacy-Preserving Case-Based Explanations. In: Fragemann, J., Li, J., Liu, X., Tsaftaris, S.A., Egger, J., Kleesiek, J. (eds) Medical Applications with Disentanglements. MAD 2022. Lecture Notes in Computer Science, vol 13823. Springer, Cham. https://doi.org/10.1007/978-3-031-25046-0_4
Download citation
DOI: https://doi.org/10.1007/978-3-031-25046-0_4
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-25045-3
Online ISBN: 978-3-031-25046-0
eBook Packages: Computer ScienceComputer Science (R0)
