Denoised Senone I-Vectors for Robust Speaker Verification

Keyphrases

• Robust Speaker Verification 100%
• I-vector 100%
• Senone 100%
• Deep Neural Network 30%
• Denoising Autoencoder 23%
• Bottleneck Features 23%
• Senone Posteriors 23%
• Gaussian Mixture Model 15%
• Vector Extraction 15%
• Feature Extraction 7%
• Extraction Methods 7%
• Recognition Task 7%
• Backpropagation 7%
• Noisy Environment 7%
• Acoustic Features 7%
• Speaker Recognition 7%
• Deep Learning Architectures 7%
• Speaker Information 7%
• NIST 2012 SRE 7%
• Phonetic Information 7%
• Posterior Estimation 7%
• Language Recognition 7%
• Deep Belief Network 7%

Computer Science

• Deep Neural Network 100%
• Speaker Verification 100%
• Denoising Autoencoder 75%
• Gaussian Mixture Model 50%
• Deep Architecture 25%
• Acoustic Feature 25%
• Extraction Process 25%
• Noisy Environment 25%
• Feature Extraction 25%

Engineering

• Deep Neural Network 100%
• Denoising Autoencoders 75%
• Gaussian Mixture Model 50%
• Feature Extraction 25%
• Extraction Process 25%
• Noisy Environment 25%
• Acoustic Feature 25%
• Deep Belief Network 25%