The Bee Colony Algorithm Optimizes The Color Image Segmentation Of Normalized Cut

Nowadays,Graph-based color image segmentation method is a hot topic of concern,in which Normalized Cut is a standardized form,which maps the image to weighted undirected graph.Specifically,the pixel that exists in the image is mapped to the node in the corresponding graph,and the weight of the edge between the connected nodes is used to represent the similarity between these pixels.However,minimizing the Normalized Cut is a NP-hard problem,many traditional methods only can get the approximate solution of the optimization problem and cannot find the optimal solution.Due to the above problems,this paper uses the segmentation method which combines the bee colony algorithm with Normalized Cut criterion,the research work is as follows:(1)This paper proposes a color image segmentation method,which is based on the Levy flight bee colony algorithm to optimize the Normalized Cut.Firstly,preprocessing the color image using FCM clustering,and then use the improved algorithm to solve the minimum value of Ncut by iterative optimization.Finally the optimal individual bee can direct the Graph partition and get the segmentation result of color image.The advantage of the Levy flight model is that it can not only enrich the population diversity,but also avoid the local optimization.The experimental results show that the proposed algorithm is more efficient and superior than other algorithms in speed convergenceand segmentation effect,such as ABC algorithm,PSO algorithm and ACO algorithm.(2)The color image segmentation method is proposed,which uses the dynamic weight of the bee colony algorithm to optimize the Normalized Cut.The first is the FCM preprocessing of color image and then the Normalized Cut criterion is optimized by the improved bee colony algorithm.The optimal individual bee is obtained to guide the image segmentation.The specific strategies are:first,the standard bee colony algorithm is discretized,given the definition of individual bee speed;and then the introduction of speed limits to rich diversity of the population;finally use dynamic weight mechanism to adjust the search mode of the individual bee according to the population evolutionary rates,balancing global search and local search capabilities.The experiments prove the effectiveness of the algorithm which is applied to color image segmentation by comparing with several other algorithms.