ISCA Archive SPSC 2021
ISCA Archive SPSC 2021

‘How to Collect Speech Data with Human Rights in Mind’ - Medical speech data analysis

Ayana Martins

Applying speech analysis for medical diagnosis requires building databases in which speech samples are paired to disease status. This is often achieved through clinical studies in collaboration with healthcare providers. In most clinical studies, data anonymization (sensu GDPR) is possible, but that is not always the case for voice recordings. Healthcare providers have high expectations on data protection and the possibility of anonymization. Thus, studies involving voice recordings require new channels and processes for collaboration. Data protection good practices have long put emphasis on security and minimizing the risk of re-identification. The GDPR has added emphasis on giving citizens a choice regarding what happens to their data. In this workshop, we highlight an example where there is a potential trade-off between these two aspects of data protection. To protect their patients, healthcare providers are normally not willing to share identifiers in the context of clinical research. How to ensure the data rights of individual citizens when we cannot identify which voice recordings belong to them?


Cite as: Martins, A. (2021) ‘How to Collect Speech Data with Human Rights in Mind’ - Medical speech data analysis. Proc. 2021 ISCA Symposium on Security and Privacy in Speech Communication,

@inproceedings{martins21_spsc,
  author={Ayana Martins},
  title={{‘How to Collect Speech Data with Human Rights in Mind’  - Medical speech data analysis}},
  year=2021,
  booktitle={Proc. 2021 ISCA Symposium on Security and Privacy in Speech Communication},
  pages={}
}