Image Threshold and Segmentation
John C. Russ · 2002
Abstract The process of reducing a gray scale or color image to a binary (black and white) image in which distinct features are isolated (usually for measurement) is called thresholding or segmentation. The words refer to slightly different but overlapping sets of procedures. Thresholding usually implies the selection of a range of gray scale or color values that distinguish the features from the background. It is widely used and very fast since it operates globally on images, but often does not provide perfect delineation of features because some pixels (particularly along feature edges) are misclassified, and features may touch each other. Segmentation implies the separation of features from the background and each other, usually based on local comparison of pixel values along feature boundaries. These methods are often slow, may not find all of the features present, and sometimes require assumptions about the shape or size of the features.