Speaker Characterization Using TDNN-LSTM Based Speaker Embedding

Keyphrases

• Neural Network 100%
• Long Short-term Memory Neural Network 100%
• Delayed Neural Networks 100%
• Speaker Recognition Evaluation 100%
• Speaker Characterization 100%
• Speaker Embedding 100%
• Decision Cost 66%
• Single System 33%
• Equal Error Rate 33%
• Training Data 33%
• Cost Function 33%
• Feature Extraction 33%
• Data Augmentation 33%
• Good Characteristics 33%
• Simple Moving Average 33%
• Evaluation Task 33%
• Indian Institute of Technology 33%
• Score Combination 33%
• National Institute of Standards 33%
• Front-end Features 33%

Computer Science

• Neural Network 100%
• time-delay 100%
• Speaker Recognition 100%
• Long Short-Term Memory Neural Network 100%
• Speaker Characterization 100%
• Experimental Result 33%
• Equal Error Rate 33%
• Training Data 33%
• Data Augmentation 33%
• Simple Average 33%
• Feature Extraction 33%