Image Colorization Based on Self-Adaptive Mutation Particle Swarm Optimization and Support Vector Machine

Image colorization technology has been widely studied because of its strong practicability. In this paper, an image colorization method of strong robustness is proposed which is based on the self-adaptive mutation particle swarm optimization and support vector machine. Firstly, the parameters of support vector machine are optimized by self-adaptive mutation particle swarm optimization. And then, the image is edited by the optimized support vector machine to separate the foreground and background. Finally, the local features of the image are reconstructed according to the based-graph theory. The experimental results show that the method in this paper can realize colorization at high fidelity. (Abstract)