Image Segmentation Based On Improved Quantum Evolutionary Kernel Clustering Algorithm

In recent years, with the rapid development of image communication technology, image segmentation has been widely applied in many areas such as pattern recognition, computer vision, medical image processing and RS vision and so on. In digital image processing, image segmentation is the key step from image processing to image analysis, and is also the basic knowledge to complete image understanding.It has emerged many classical methods and theories in the field of image segmentation algorithm, in which Fuzzy C-Means algorithm is widely used, but has difficult to overcome its own shortcomings. For example, it is sensitive to noise, and the clustering effect of quality is related to the initial sample. Therefore, on the basis of traditional fuzzy C-means clustering algorithm, improvements are given as follows:(1) Kernel Function is adopted to improve the traditional fuzzy C-means clustering algorithm, which is achieved by using kernel distance to build objective function. Further, another improvement is achieved by incorporates a penalty term, which controls the neighborhood, to the objective function. It can be seen from the simulation results that the improved kernel clustering algorithm can not only suppress noise effectively, but also can improve segmentation effect of noise image.(2) The angle of the rotation gate in the traditional algorithm was modified according to the evolutionary generation and the fitness values, which is based on the simulated annealing idea. The stability and accuracy of the algorithm are both improved.(3) The improved quantum evolutionary algorithm and improved kernel clustering algorithm are combined to form the improved quantum evolutionary kernel clustering algorithm, at the same time this algorithm was applied to segment image.It can be seen from the simulation results that the algorithm, which is proposed in this paper, can not only robust to noise, but can overcome the characteristics of traditional evolutionary algorithm effectively, such as slow convergence and early maturity easily. In addition, the accuracy and stability of algorithm could be improved greatly.