Hiding Speaker's Sex in Speech Using Zero-Evidence Speaker Representation in an Analysis/Synthesis Pipeline.
Paul-Gauthier NoéXiaoxiao MiaoXin WangJunichi YamagishiJean-François BonastreDriss MatroufPublished in: ICASSP (2023)
Keyphrases
- speech recognition
- audio visual
- speaker recognition
- speaker verification
- automatic speech recognition
- speaker identification
- speaker dependent
- neural network
- speaker diarization
- structural analysis
- hidden markov models
- multiscale
- texture synthesis
- visual features
- statistical analysis
- image analysis
- speaker independent
- prosodic features
- speaker adaptation