Selective HuBERT: Self-Supervised Pre-Training for Target Speaker in Clean and Mixture Speech.
Jingru LinMeng GeWupeng WangHaizhou LiMengling FengPublished in: IEEE Signal Process. Lett. (2024)
Keyphrases
- speech recognition
- speaker recognition
- acoustic models
- audio visual
- automatic speech recognition
- speaker identification
- speaker verification
- gaussian mixture model
- hearing impaired
- mixture model
- speaker diarization
- hidden markov models
- speaker dependent
- automatic speech recognition systems
- speech signal
- training algorithm
- training phase
- training set
- vocal tract
- audio stream
- noisy speech
- moving target
- automatic transcription
- prosodic features
- learning algorithm
- broadcast news
- speech sounds
- supervised learning
- training examples
- acoustic features
- probabilistic neural network