All Published Papers Search Service

Title

A fast multi-level image segmentation method based on 2-D histogram using PSO

Author

Zeng Guowen, Hao Zhifeng, Yang Xiaowei, Wu Guangchao

Vol. 6  No. 1  pp. 126~135

Abstract

In order to cope with the problem that 2-D thresholding method can not deal with multi-level image segmentation problem properly, we present one new method for dividing the 2-D histogram plane into multi-object areas, then deal with the problem in each sub-area, at last gather the information together to find the solution. In order to decrease calculation time of the 2-D thresholding method, we present a fast computation technique with which simple algebraic sum can replace the accumulated value. In the whole we use particle swarm optimization algorithm to seek the global optimal solution. We make it possible that the 2-D thresholding technique can solve the multi-level image segmentation problems quickly. The experimental results prove that the proposal method can deal with arbitrary multi-level complex image segmentation problem efficiently. The behavior of the proposal method is better than multi-level image segmentation algorithm based on 1-D histogram and some other multi-level thresholding methods. We also prove that PSO is a good way to solve the multi-level image segmentation problem.

Keywords

Multi-level thresholding, particle swarm optimization (PSO), 2-D threshold

URL