Acoustic analysis of hypokinetic dysarthria (HD), which is present in approximately 90% of patients with Parkinson’s disease (PD), is quick and effective technique that can identify PD in first stages and thus can indirectly keep patient’s quality of live and decrease the costs related to medication. However, engineers around the world are facing a big problem and that is the lack of databases. The procedure of PD patients acquisition is very time consuming and expensive. Moreover, scientists cannot share the databases due to speakers’ identity. One way how to disseminate these databases and fasten the process of HD research is speakers’ de-identification.
De-identification of PD patients is not trivial, because on one hand we must de-identify speaker, but on the other hand we must keep the information about speaker’s voice pathology. In the frame of COST-STSM-IC1206-32625 we proposed a new concept of dysarthric speech de-identification that is able to keep this information and that brings new possibilities of dysarthric speech corpora dissemination.