Unsupervised learning of HMM topology for text-dependent speaker verification

Keyphrases

• Acoustic Matching 33%
• Adaptation 33%
• Adaptation Methods 33%
• Adaptive Adjustment 33%
• Background Modeling 33%
• Best System 33%
• Dependent Systems 33%
• Enrollment Data 66%
• Expectation Maximization 33%
• Expectation-maximization Algorithm 33%
• Generic Model 33%
• HMM Topology 100%
• Impostor 33%
• Label Information 33%
• Moderate Training 33%
• Phonetics 33%
• Popular 33%
• Specific Training 33%
• Target Speaker 33%
• Text-dependent 66%
• Text-independent Speaker Verification 100%
• Topological Information 33%
• Training Data 100%
• Unsupervised Learning 100%
• Unsupervised Learning Techniques 33%
• Utterance 66%

Computer Science

• background modeling 33%
• Enrollment Data 66%
• Experimental Result 33%
• Phonetic Content 33%
• Speaker Verification 100%
• Traditional Method 33%
• Training Data 100%
• Unsupervised Learning 100%