Multi-frame feature-fusion-based model for violence detection

Keyphrases

• Spatial Features 100%
• Surveillance System 100%
• Violence Detection 100%
• Feature Fusion 100%
• Feature Fusion Method 100%
• Multi-frame Feature 100%
• Residual Dense Block 100%
• Human Behavior 50%
• Concatenated 50%
• State-of-the-art Techniques 50%
• Convolutional Neural Network 50%
• Superior Performance 50%
• Two-input 50%
• Publicly Available Dataset 50%
• Temporal Features 50%
• Feature Combination 50%
• Learned Features 50%
• Multi-level Features 50%
• Public Safety 50%
• Human-based 50%
• Public Monitoring 50%
• Human Attention 50%
• Human Observation 50%
• Temporal Dependence 50%
• Violent Behavior 50%
• Long Short-term Memory Units 50%
• Autonomous Surveillance 50%
• Motion Information 50%
• Video Surveillance 50%
• Human Detection 50%
• Action-based 50%

Computer Science

• Feature Fusion 100%
• Experimental Result 33%
• Convolutional Neural Network 33%
• Superior Performance 33%
• Long Short-Term Memory Network 33%
• Temporal Feature 33%
• Learned Feature 33%
• Motion Information 33%
• Video Surveillance 33%