An image threholding approach based on cuckoo search algorithm and 2D maximum entropy

The image thresholding approach based on the basis of 2-D maximum entropy has better segmentation performance by the use of local space information of pixels, but it is unpractical for heavy computation required by this method. In the paper, an image segmentation technology based on cuckoo search and 2-D maximum entropy is presented, which views the seeking of 2-D maximum entropy of the image as a function optimization problem and uses the behavior of the obligate brood parasitism of some cuckoo species to simulate the process of searching optimal threshold. Furthermore, a local search strategy is employed to improve the results in the cuckoo search algorithm. The experimental results proves that compared with 2-D maximum entropy thresholding optimized with genetic algorithm, differential evolution algorithm and particle swarm optimization algorithm, the proposed method is able to get the optimal thresholds quickly and with a higher probability to get optimal threshold, which is a fast and robust image segmentation method.