A Gaussian-mixture based approach to spatial image background modeling and compensation.
Afsar SaranliPublished in: EUSIPCO (2007)
Keyphrases
- gaussian mixture
- mixture of gaussians
- spatial information
- background modeling
- image features
- input image
- template matching
- multiscale
- complex background
- gaussian mixture model
- varying illumination
- image classification
- probability density function
- mixture model
- pixel values
- image regions
- closed form
- motion estimation
- density estimation
- high resolution
- image representation
- object detection
- lighting conditions
- object recognition
- pixel level
- low resolution
- unsupervised learning
- image segmentation