Title
Analysis of gender and identity issues in depression detection on de-identified speech
Abstract
•Depression detection tools can help to remotely monitor the patients’ mental state.•Speaker de-identification allows the protection of patients’ privacy.•Depression detection with different speaker de-identification techniques is evaluated.•Gender-dependent de-identification leads to better depression detection results.•Speaker-dependent de-identification do not outperform speaker-independent approaches.
Year
DOI
Venue
2021
10.1016/j.csl.2020.101118
Computer Speech & Language
Keywords
DocType
Volume
Speaker de-identification,Depression detection,Voice conversion,Frequency warping and amplitude scaling,Generative adversarial networks
Journal
65
ISSN
Citations 
PageRank 
0885-2308
0
0.34
References 
Authors
0
2
Name
Order
Citations
PageRank
Paula Lopez-Otero16413.18
Laura Docio-Fernandez2204.00