Avoiding Speaker Overfitting in End-to-End DNNs Using Raw Waveform for Text-Independent Speaker Verification.
Jee-weon JungHee-Soo HeoIl-Ho YangHye-jin ShimHa-Jin YuPublished in: INTERSPEECH (2018)
Keyphrases
- end to end
- speaker verification
- speaker recognition
- noisy environments
- prosodic features
- language identification
- audio visual
- congestion control
- acoustic features
- emotion recognition
- information retrieval
- multilayer perceptron
- text to speech
- using artificial neural networks
- face verification
- edge detection
- keywords
- high level
- multi modal
- speaker identification
- text mining
- low level
- speaker diarization
- neural network