A fast two-dimensional entropic thresholding algorithm

摘要

Two-dimensional (2D) entropic thresholding is one of the important thresholding techniques for image segmentation. The selection of the global threshold vector is usually through a “maximin” optimization procedure. A fast two-phase 2D entropic thresholding algorithm is proposed. In order to reduce the computation time, first 9L23 candidate threshold vectors are estimated from a quantized image of the original. The global threshold vector is then obtained by checking candidates only. The optimal computation complexity is O(L23) by quantizing the gray level into L23 levels. Experimental results show that the processing time of each image is reduced from more than 2 h to about 2 min. The required memory space is also greatly reduced.