Optimal feature selection based speech emotion recognition using two-stream deep convolutional neural network

Keyphrases

• Active Area 33%
• Behavior Assessment 33%
• Convolutional Neural Network 33%
• Deep Convolutional Neural Network (deep CNN) 100%
• Discriminative Information 33%
• Dominant Source 33%
• Emo-DB 33%
• Emotion Recognition 33%
• Emotional Corpora 33%
• Feature Extraction 66%
• Feature Learning 33%
• Final Model 33%
• Fully Connected Network 33%
• Human Behavior 33%
• Human Being 33%
• Human-computer Interaction 66%
• Information Exchange 33%
• Interaction Zone 33%
• Iterative Neighborhood Component Analysis 66%
• Model Training 33%
• Neural Network Structure 33%
• Optimal Feature Selection 100%
• Performance Prediction 33%
• RAVDESS 33%
• Recognition Rate 33%
• Saliency Information 33%
• SAVEE 33%
• Severance 33%
• Softmax Classifier 33%
• Spatial Domain 33%
• Spectral Domain 33%
• Spectral-spatial Feature 33%
• Speech Emotion Recognition 100%
• Speech Emotion Recognition Systems 33%
• Speech Processing 33%
• Speech Signal 33%
• Two-channel 33%
• Two-stream 100%

Computer Science

• Component Analysis 66%
• Connected Network 33%
• Convolutional Neural Network 33%
• Deep Convolutional Neural Networks 100%
• Deep Feature 33%
• Feature Extraction 100%
• Feature Selection 100%
• Neural Network Architecture 33%
• Prediction Performance 33%
• Recognition Rate 33%
• Recognition System 33%
• Representation Learning 33%
• Signal Processing 33%
• Spatial Domain 33%
• Spectral Domain 33%
• Spectral Feature 33%
• speech corpus 33%
• Speech Emotion Recognition 100%