Image segmentation Based on Chaotic Quantum Ant Colony Algorithm

Abstract

Ant colony algorithm is a new type of biomimetic evolutionary algorithm, which has some outstanding characteristics like good robustness, parallelism and positive feedback. It has been widely used in many fields but it still has some weaknesses, such as slow-convergence and easily-falling into local extreme value. Therefore we propose a new algorithm which combines the quantum evolutionary algorithm and ant colony algorithm together based on these shortcomings, this new type of algorithm considers the two quantum bit probability amplitudes as ants’ current location information. The searching space will be doubled when ants’ numbers remain the same. Both introducing the pixel point gradient into quantum revolving door, and dynamically changing rotation angle are to achieve local-searching. Searching by chaotic quanta near candidates with optimal solution is to improve the global optimization. Meanwhile, adopting a NAND gate is to achieve mutating operation, and to avoid algorithm’s precocity. Compared with the traditional algorithm, the new algorithm has a better population diversity and an outstanding ability to overcome the precocity and stagnation in the optimization process. It has been proved that this improved algorithm is effectively to the problems like slow-convergence and easily-falling into local extreme value, and vividly increase the speed and accuracy of the image segmentation.