Zero-Shot Unseen Speaker Anonymization via Voice Conversion.
Hyung-Pil ChangIn-Chul YooChanghyeon JeongDongsuk YookPublished in: IEEE Access (2022)
Keyphrases
- synthesized speech
- prosodic features
- information loss
- privacy preserving
- speech sounds
- speaker recognition
- mel frequency cepstral coefficients
- speaker verification
- privacy preservation
- speaker identification
- speech recognition
- audio visual
- text to speech
- data publishing
- training examples
- previously unseen
- emotion recognition
- training set
- privacy preserving data publishing
- speaker diarization
- privacy protection
- automatic speech recognition
- speaker dependent
- speech synthesis
- object categories
- vocal tract
- neural network
- preserving privacy
- data privacy
- data anonymization
- hidden markov models