Weakly Supervised U-Net with Limited Upsampling for Sound Event Detection

Mathematics

• Training Data 100%
• Convolutional Layer 100%
• Sound Source 100%
• Connected Layer 100%
• Global Method 100%

Computer Science

• Event Detection 100%
• U-Net 100%
• Global Threshold 37%
• Detection Performance 25%
• Anomaly Detection 12%
• Experimental Result 12%
• Supervised Learning 12%
• Performance Degradation 12%
• Timing Information 12%
• Training Data 12%
• Computation Time 12%
• Convolutional Layer 12%
• timing offset 12%
• Audio Recording 12%

Engineering

• Event Detection 100%
• Tasks 33%
• Detection Performance 33%
• Experimental Result 16%
• Convolutional Layer 16%
• Performance Degradation 16%
• Computation Time 16%
• Sound Source 16%
• Surveillance System 16%
• Detection Task 16%
• Anomaly Detection 16%
• Audio Recording 16%
• Global Method 16%
• Output Map 16%
• Frequency Axis 16%