Anomaly Detection in Video Sequence with Appearance-Motion Correspondence.
Trong-Nguyen NguyenJean MeunierPublished in: CoRR (2019)
Keyphrases
- anomaly detection
- video sequences
- motion estimation
- camera motion
- space time
- image sequences
- dynamic textures
- intrusion detection
- moving objects
- object motion
- detecting anomalies
- adjacent frames
- human motion
- anomalous behavior
- optical flow
- video objects
- motion field
- network traffic
- key frames
- network intrusion detection
- unsupervised anomaly detection
- network anomaly detection
- intrusion detection system
- motion vectors
- dynamic scenes
- one class support vector machines
- feature points
- network security
- video data
- behavior analysis
- detecting anomalous
- motion model
- point correspondences
- motion compensation
- malware detection
- pattern recognition
- appearance model
- unsupervised learning
- video frames
- machine learning
- negative selection algorithm
- training set
- feature vectors
- maximum likelihood
- model selection
- maximum entropy