Optimal multi-level thresholding with membrane computing

Abstract

The conventional methods are not effective and efficient for image multi-level thresholding due to time-consuming and expensive computation cost. The multi-level thresholding problem can be posed as an optimization problem, optimizing some thresholding criterion. In this paper, membrane computing is introduced to propose an efficient and robust multi-level thresholding method, where a cell-like P system with the nested structure of three layers is designed as its computing framework. Moreover, an improved velocity-position model is developed to evolve the objects in membranes based on the special membrane structure and communication mechanism of objects. Under the control of evolution-communication mechanism of objects, the cell-like P system can efficiently exploit the best multi-level thresholds for an image. Simulation experiments on nine standard images compare the proposed multi-level thresholding method with several state-of-the-art multi-level thresholding methods and demonstrate its superiority.