Houjun Huang

Latest

• Unit Selection Synthesis based Data Augmentation for Fixed Phrase Speaker Verification
• Margin matters: Towards more discriminative deep neural network embeddings for speaker recognition