A multi-level thresholding image segmentation based on an improved artificial bee colony algorithm

As a popular evolutionary algorithm, Artificial Bee Colony (ABC) algorithm has been successfully applied into threshold-based image segmentation. Due to its one dimension search strategy, the convergence speed of ABC is slow and its solution is acceptable but not precise. For making more fine-tuning search and further enhancing the achievements on image segmentation, we proposed an Otsu segmentation method based on a new ABC algorithm. Different from the traditional ABC strategy, our algorithm takes full use of individuals information which is defined by a focus point and the best point to increase its accuracy and convergence speed. Furtheremore, we propose an adaptive parameter to adjust the search step of individual automatically, which also improves its exploitation ability. Experimental results on Berkeley segmentation database demonstrate the effectiveness of our algorithm.