A novel discriminative feature extraction for acoustic scene classification using RNN based source separation

Keyphrases

• Neural Network 100%
• Recurrent Neural Network 100%
• Discriminative Feature Extraction 100%
• Source Separation 100%
• Acoustic Scene Classification 100%
• Feature Extraction Methods 50%
• Acoustic Event 50%
• Acoustic Scene 50%
• Similarity Classes 50%
• Low Performance 25%
• Discriminative Features 25%
• Classification Accuracy 25%
• Discriminative Information 25%
• Performance Problems 25%
• Performance Limitation 25%
• Improvement Factor 25%
• Deep Learning Features 25%
• Sound System 25%
• Accuracy Performance 25%
• Neural Network Structure 25%
• Class Prior 25%
• Human Ear 25%
• Final Evaluation 25%
• Top 10 25%

Computer Science

• Discriminative Feature 100%
• Recurrent Neural Network 100%
• Source Separation 100%
• Feature Extraction 100%
• Lower Performance 25%
• Conventional Method 25%
• Layer Information 25%
• Performance Problem 25%
• Neural Network Architecture 25%
• Deep Learning Method 25%

Engineering

• Feature Extraction 100%
• Source Separation 100%
• Recurrent Neural Network 100%
• Isolation Method 50%
• Limitations 25%
• Conventional Method 25%
• Layer Information 25%
• Based Feature Extraction 25%
• Human Ear 25%
• Deep Learning Method 25%