Improved hybrid particle swarm optimisation for image segmentation

Abstract

A method for image segmentation based on improved hybrid particle swarm optimisation (PSO) is proposed. In view of the shortcoming that the traditional PSO algorithm is easy to fall into local optimal solution, we update the particle velocity based on the combination of global optimisation, region equilibrium and compression factor. By this way, the searchability of the particle and optimisation performance of the improved PSO is improved. Experiments results on three classic test functions show that the algorithm can greatly improve the searchability. Experiments also show that it performs well on image segmentation.

GRAPHICAL ABSTRACT

KEYWORDS:

• Hybrid particle swarm optimisation
• region equilibrium
• compression factor
• image segmentation

Disclosure statement

No potential conflict of interest for all authors.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China [grant numbers 61179032 and 61303116].