Double-constrained RPCA based on saliency maps for foreground detection in automated maritime surveillance.
Andrews SobralThierry BouwmansEl-hadi ZahzahPublished in: AVSS (2015)
Keyphrases
- foreground detection
- saliency map
- video surveillance
- surveillance system
- human detection
- visual attention
- background subtraction
- foreground objects
- object segmentation
- background modeling
- robust principal component analysis
- human visual system
- computer vision
- dynamic background
- moving object detection
- video frames
- input image
- natural images
- motion detection
- event detection
- moving objects
- low level features
- real time
- dynamic scenes
- visual features
- image sequences
- foreground segmentation
- gaussian mixture model