The Chinese Dysarthria Speech Database (CDSD) is a valuable resource for dysarthria research. It contains 34 hours of speech data from 24 participants with dysarthria. The data was recorded using mobile devices in a natural environment, enhancing its ecological validity. The paper also introduces the concept of Speaker-dependent Speech Recognition, which offers a solution to address speech recognition challenges among dysarthria speakers.
Publication date: 24 Oct 2023
Project Page: https://arxiv.org/abs/2310.15930v1
Paper: https://arxiv.org/pdf/2310.15930
