A pixel-level, intensity-based nonlinear autoregressive classifier (NARX) with chromatic exogenous input for efficient image background subtraction.
Syed A. YusufDavid J. BrownAlan MackinnonRichard PapanicolaouPublished in: IJCNN (2013)
Keyphrases
- pixel level
- foreground objects
- autoregressive
- background subtraction
- background model
- moving objects
- template matching
- video surveillance
- video sequences
- object segmentation
- object detection
- superpixels
- image fusion
- non stationary
- random fields
- image sequences
- input image
- dynamic scenes
- gaussian mixture model
- segmentation method
- image analysis
- moving camera
- computer vision
- post processing
- appearance model
- edge detection
- image features
- multiscale
- similarity measure
- high resolution
- video data
- sar images
- conditional random fields
- maximum likelihood
- image registration
- multi class
- optical flow
- object recognition
- image segmentation