Unsupervised Video Anomaly Detection via Flow-based Generative Modeling on Appearance and Motion Latent Features.
MyeongAh ChoTaeoh KimSangyoun LeePublished in: CoRR (2020)
Keyphrases
- anomaly detection
- appearance features
- static images
- unsupervised learning
- intrusion detection
- unsupervised anomaly detection
- key frames
- motion features
- intrusion detection system
- object motion
- network intrusion detection
- detecting anomalies
- anomalous behavior
- dynamic textures
- video sequences
- behavior analysis
- network traffic
- feature vectors
- image sequences
- one class support vector machines
- probabilistic topic models
- generative model
- supervised learning
- motion estimation
- action recognition
- detect anomalies
- feature extraction
- semi supervised
- optical flow
- malware detection
- motion model
- moving objects
- pairwise
- video frames
- network anomaly detection
- negative selection algorithm
- gaussian process latent variable models