Criterion for Automatic Selection of the Most Suitable Maximum-Likelihood Thresholding Algorithm for Extracting Object from their Background in a Still Image.
Geovanni MartinezPublished in: MVA (2005)
Keyphrases
- automatic selection
- thresholding algorithm
- maximum likelihood
- bounding box
- background clutter
- foreground and background
- image data
- image regions
- background noise
- lighting conditions
- image features
- input image
- multiple objects
- image representation
- single image
- pixel level
- background pixels
- multiscale
- complex scenes
- image segmentation
- criterion function
- spatial relationships
- normalized correlation
- d objects
- image classification
- template matching
- complex background
- object contours
- salient objects
- cluttered background
- keypoints
- object class
- target object
- partial occlusion
- image retrieval
- image content
- spatial information
- color distribution
- test images
- object shapes
- object appearance
- edge detection
- segmentation algorithm
- object regions
- figure ground segmentation
- object region
- computer vision
- high resolution
- image segments
- segmentation method
- object model
- pixel values
- region of interest
- spatial relations