Anomaly3D: Video anomaly detection based on 3D-normality clusters

Keyphrases

• Normality 100%
• Video Anomaly Detection 100%
• Abnormal Detection 50%
• 3D Convolutional Autoencoder 50%
• Performance Improvement 25%
• Benchmark Dataset 25%
• Labeled Data 25%
• Surveillance System 25%
• Vision Impairment 25%
• SVM Classifier 25%
• Unsupervised Method 25%
• Event-based 25%
• Training Video 25%
• Public Safety 25%
• One-class Support Vector Machine (OCSVM) 25%
• Surveillance Video 25%
• Strong Pattern 25%
• 3D Convolution 25%
• Motion Features 25%
• Human-based 25%
• Encoder Architecture 25%
• Anomalous Behavior 25%
• Abnormal Events 25%
• Public Monitoring 25%
• Human Attention 25%
• Two-stage Architecture 25%
• Autonomous Detection 25%
• Autoencoder Architecture 25%
• Encoder-decoder Architecture 25%
• Spatio-temporal Features 25%
• 3D Spatiotemporal Features 25%
• Human Observation 25%

Computer Science

• Temporal Feature 100%
• Autoencoder 100%
• Anomaly-Based Detection 100%
• Experimental Result 50%
• Support Vector Machine 50%
• Performance Improvement 50%
• Event Detection 50%
• Unsupervised Manner 50%
• surveillance video 50%
• Video Training 50%
• Anomalous Behavior 50%
• Motion Feature 50%