Optimized discriminative transformations for speech features based on minimum classification error

Keyphrases

• Feature-based 100%
• Data Dependency 100%
• Speech Features 100%
• Minimum Classification Error 100%
• Hidden Markov Model 66%
• Feature Transformation 66%
• Heteroscedastic 66%
• Classification Error 66%
• PCA-LDA 66%
• Performance Improvement 33%
• Feature Extraction 33%
• Classification Performance 33%
• Environmental Conditions 33%
• TIMIT 33%
• Speech Recognition 33%
• Feature Vector 33%
• Dependent Features 33%
• Recognition Task 33%
• Optimization Criterion 33%
• Pattern Classification 33%
• Transformation Matrix 33%
• Linear Discriminant Analysis 33%
• AURORA2 33%
• Phone Recognition 33%
• Isolated Word Recognition 33%
• Classification Recognition 33%
• Mel-frequency Cepstral Coefficients 33%

Computer Science

• Classification Error 100%
• Minimum Classification 100%
• Linear Discriminant Analysis 80%
• Component Analysis 40%
• Principal Components 40%
• Experimental Result 20%
• Feature Vector 20%
• Speech Recognition 20%
• Optimization Criterion 20%
• Transformation Matrix 20%
• Extracted Feature 20%
• Pattern Classification 20%
• Feature Extraction 20%
• Mel-Frequency Cepstral Coefficients 20%