Histograms of Optical Flow Orientation and Magnitude and Entropy to Detect Anomalous Events in Videos.
Rensso Victor Hugo Mora ColqueCarlos CaetanoMatheus Toledo Lustosa de AndradeWilliam Robson SchwartzPublished in: IEEE Trans. Circuits Syst. Video Technol. (2017)
Keyphrases
- optical flow
- event recognition
- normal behavior
- human activities
- video event
- detecting anomalous
- anomaly detection
- event detection
- moving camera
- detect anomalies
- sports video
- video analysis
- video clips
- image sequences
- orientation histogram
- motion features
- video sequences
- temporal relationships
- spatio temporal patterns
- detection method
- optical flow estimation
- computer vision
- information theory
- surveillance videos
- image brightness
- optical flow computation
- motion estimation
- intrusion detection
- detection algorithm
- video dataset
- video data
- motion field
- action recognition
- video surveillance
- moving objects
- independently moving objects
- spatio temporal
- motion model
- information theoretic
- activity recognition
- temporal information
- motion segmentation
- video content
- temporal patterns
- mutual information
- feature tracking
- unusual events
- motion trajectories
- dense optical flow
- dynamic textures
- camera motion
- d scene
- lifelog
- video segments
- human actions
- video shots