Image Thresholding Using a Membrane Algorithm Based on Enhanced Particle Swarm Optimization with Hyperparameter
Dequan Guo, Gexiang Zhang, Yi Zhou, Jianying Yuan, Prithwineel Paul, Kechang Fu and Ming Zhu
Image thresholding is an important research direction of image segmentation that aims to divide image into meaningful sub-regions. This paper introduces an optimal thresholding by a cell-like membrane algorithm with enhanced particle swarm optimization (PSO) with hyperparameter, namely MAPSOH. Under the membrane evolution-communication mechanism, the designed hyperparameter method for PSO parameters can obtain better convergence in less time. According to the special membrane structure, a modification of PSO is employed to find the best multi-level thresholding for image segmentation problem effectively. The experiments demonstrate that the proposed method has better quantitative statistical comparisons and qualitative performance in comparison with several existing methods.
Keywords: Image segmentation, particle swarm optimization, membrane computing, P systems, hyperparameter, thresholding approach