Image Thresholding Based on Improved 2-D Minimum Within-Cluster Absolute Difference Method and Its Fast Algorithm

In view of the inaccurate segmented result and low computational efficiency of the existing thresholding method based on minimum withincluster absolute difference, an improved thresholding method based on recursive chaotic particle swarm algorithm is proposed in this paper. To achieve higher segmented accuracy, the gray scale-gradient histogram is introduced firstly, and then the formulae for the proposed method are simplified and the corresponding fast recursive method is presented. After that, the improved Tent map chaotic particle swarm algorithm is used to search for the optimal threshold, and the repeat computations of the fitness function in iteration are reduced significantly using recursive mode Compared with the exhaustive algorithm based on gray scale-average gray scale histogram and minimum within-cluster absolute difference, experimental results show that the new method obtain more accurate threshold and achieve better segmented result since the effect of the edge points and noise points are removed. And because the swarm intelligence is used to optimize the searching process, the running time is reduced by two orders. Compared with method of maximum between-cluster variance based on gray scale-gradient histogram and logistic map recursive chaotic particle swarm algorithm, the new method is based on improved Tent map, higher convergence precision is obtained because of its higher ergodicity