Short-time spectral aggregation for speaker embedding

Keyphrases

• Time-spectral 100%
• Speaker Embedding 100%
• Spectral Aggregation 100%
• Spectral Pooling 83%
• Feature Map 50%
• Spectral Components 50%
• Spectrogram 33%
• Statistical Pooling 33%
• Frame-level 33%
• High Performance 16%
• Spectral Domain 16%
• Multi-head 16%
• Temporal Features 16%
• Spectral Information 16%
• Speaker Recognition System 16%
• Acoustic Features 16%
• Speaker Information 16%
• VoxCeleb 16%
• Frame-level Features 16%
• Attention Pooling 16%
• CNN Features 16%
• Embedded Networks 16%
• Utterance Level 16%
• Fixed-dimensional Embedding 16%

Computer Science

• Spectral Component 100%
• Feature Map 100%
• Convolutional Neural Network 33%
• Verification System 33%
• Speaker Verification 33%
• Level Representation 33%
• Temporal Feature 33%
• Spectral Domain 33%
• Acoustic Feature 33%
• Spectral Information 33%