A fast recurring two-dimensional entropic thresholding algorithm

Thresholding is an important form of image segmentation and is used in the processing of images for many applications. One of the criteria to select a suitable threshold is the maximization of the two-dimensional (2-D) entropies based on the 2-D (gray-level/local average gray-level) histogram. The rationale of this approach is introduced. In order to reduce the computation time of entropy function, a fast recurring algorithm for 2-D entropic thresholding method is presented. The experimental results show that the processing time to obtain the threshold vector from 2-D histogram is reduced from 30 to 0.15 s. (C) 1999 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.