Generative Modeling of Biomedical Audio Signals for Respiratory Disease Screening Using Reference-Guided Feature Adaptation

Authors

  • Kristopher Carney Department of Computer Science, University of Houston, Houston, TX, USA.
  • Meike Stolze Department of Computer Science, Colorado State University, Fort Collins, CO, USA.

Keywords:

respiratory audio; generative modeling; reference-guided adaptation; domain shift; biomedical signal processing; uncertainty-aware screening

Abstract

Respiratory audio offers a low-cost channel for screening conditions such as asthma, chronic obstructive pulmonary disease, pneumonia, and viral respiratory infection, but learned screeners are fragile under acquisition and population shift. Microphone response, room acoustics, breathing effort, body habitus, and subject-specific timbre can dominate the subtle acoustic cues associated with disease. Generative models could improve robustness by learning latent structure and synthesizing underrepresented conditions, yet unconstrained synthesis can reproduce device artifacts, speaker identity, and other nuisance factors rather than pathology. This manuscript develops a reference-guided generative feature adaptation framework for respiratory disease screening. A short, label-free reference recording conditions a scale-shift transformation in a spectral latent space, while a conditional generative model reconstructs and samples respiratory spectro-temporal representations. A screening classifier operates on adapted latents and is regularized by disease classification, domain-adversarial, reference-consistency, and reconstruction objectives. The framework separates what the reference is allowed to explain from what the classifier is allowed to use, and it couples predictions with calibration and uncertainty gates. Analytical assessment identifies the central trade-off: adaptation strong enough to remove device and subject style can also erase disease-related transients, while weak adaptation leaves covariate shift intact. The contribution is a technically specified, testable design that links reference conditioning, generative augmentation, nuisance reduction, and uncertainty-aware decision rules. It provides a foundation for rigorous subject-level and device-held-out evaluation of respiratory audio screening before clinical deployment.

References

1. Brown, C., Chauhan, J., Grammenos, A., Han, J., Hasthanasombat, A., Spathis, D., Xia, T., Cicuta, P., & Mascolo, C. (2020). Exploring automatic diagnosis of COVID-19 from crowdsourced respiratory sound data. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (pp. 3474–3484). Association for Computing Machinery. https://doi.org/10.1145/3394486.3412865

2. Pahar, M., Klopper, M., Warren, R., & Niesler, T. (2021). COVID-19 cough classification using machine learning and global smartphone recordings. Computers in Biology and Medicine, 135, 104572. https://doi.org/10.1016/j.compbiomed.2021.104572

3. Sharma, N., Krishnan, P., Kumar, R., Ramoji, S., Chetupalli, S. R., Ghosh, P. K., & Ganapathy, S. (2020). Coswara—A database of breathing, cough, and voice sounds for COVID-19 diagnosis. arXiv. https://arxiv.org/abs/2005.10548

4. Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., & Bengio, Y. (2014). Generative adversarial nets. In Advances in Neural Information Processing Systems (Vol. 27, pp. 2672–2680). Curran Associates.

5. Kingma, D. P., & Welling, M. (2014). Auto-encoding variational Bayes. In International Conference on Learning Representations.

6. Ho, J., Jain, A., & Abbeel, P. (2020). Denoising diffusion probabilistic models. In Advances in Neural Information Processing Systems (Vol. 33, pp. 6840–6851). Curran Associates.

7. Oord, A. v. d., Dieleman, S., Zen, H., Simonyan, K., Vinyals, O., Graves, A., Kalchbrenner, N., Senior, A., & Kavukcuoglu, K. (2016). WaveNet: A generative model for raw audio. arXiv. https://arxiv.org/abs/1609.03499

8. Kong, J., Kim, J., & Bae, J. (2020). HiFi-GAN: Generative adversarial networks for efficient and high fidelity speech synthesis. In Advances in Neural Information Processing Systems (Vol. 33, pp. 17022–17033). Curran Associates.

9. Ganin, Y., Ustinova, E., Ajakan, H., Germain, P., Larochelle, H., Laviolette, F., Marchand, M., & Lempitsky, V. (2016). Domain-adversarial training of neural networks. Journal of Machine Learning Research, 17(59), 1–35.

10. Sun, B., Feng, J., & Saenko, K. (2016). Return of frustratingly easy domain adaptation. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 30, No. 1, pp. 2058–2065).

11. Finlayson, S. G., Subbaswamy, A., Singh, K., Bowers, J., Kupke, A., Zittrain, J., Kohane, I. S., & Saria, S. (2021). The clinician and dataset shift in artificial intelligence. New England Journal of Medicine, 385(3), 283–286. https://doi.org/10.1056/NEJMc2104626

12. Seyyed-Kalantari, L., Zhang, H., McDermott, M. B., Chen, I. Y., & Ghassemi, M. (2021). Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations. Nature Medicine, 27(12), 2176–2182. https://doi.org/10.1038/s41591-021-01595-0

13. Kiyasseh, D., Zhu, T., & Clifton, D. A. (2021). CLOCS: Contrastive learning of cardiac signals across space, time, and patients. In Proceedings of the 38th International Conference on Machine Learning (Vol. 139, pp. 5606–5615). PMLR.

14. Zhang, Y., Huang, R., Li, R., He, J., Xia, Y., Chen, F., ... & Zhao, Z. (2024, March). Stylesinger: Style transfer for out-of-domain singing voice synthesis. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 38, No. 17, pp. 19597-19605).

15. Perez, E., Strub, F., de Vries, H., Dumoulin, V., & Courville, A. (2018). FiLM: Visual reasoning with a general conditioning layer. In Proceedings of the AAAI Conference on Artificial Intelligence (Vol. 32, No. 1, pp. 3942–3951).

16. Karras, T., Laine, S., & Aila, T. (2019). A style-based generator architecture for generative adversarial networks. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (pp. 4401–4410).

17. Gal, Y., & Ghahramani, Z. (2016). Dropout as a Bayesian approximation: Representing model uncertainty in deep learning. In Proceedings of the 33rd International Conference on Machine Learning (Vol. 48, pp. 1050–1059). PMLR.

18. Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017). On calibration of modern neural networks. In Proceedings of the 34th International Conference on Machine Learning (Vol. 70, pp. 1321–1330). PMLR.

19. Roberts, M., Driggs, D., Thorpe, M., Gilbey, J., Yeung, M., Ursprung, S., Aviles-Rivero, A. I., Etmann, C., McCague, C., Beer, L., Weir-McCall, J. R., Teng, Z., Gkrania-Klotsas, E., Rudd, J. H. F., Sala, E., & Schönlieb, C.-B. (2021). Common pitfalls and recommendations for using machine learning to detect and prognosticate for COVID-19 using chest radiographs and CT scans. Nature Machine Intelligence, 3(3), 199–217. https://doi.org/10.1038/s42256-021-00307-1

Downloads

Published

2026-11-06

How to Cite

Kristopher Carney, & Meike Stolze. (2026). Generative Modeling of Biomedical Audio Signals for Respiratory Disease Screening Using Reference-Guided Feature Adaptation. Bioinformatics Insights and Analytics, 4(3). Retrieved from https://bioinfia.org/index.php/home/article/view/246