An effective method for image segmentation

This paper presents an adaptive immune genetic algorithm (AIGA) for image segmentation based on the cost minimization technique. The image segmentation problem is treated as one of combinatorial optimization. A cost function which incorporates both edge information and region gray-scale uniformity is used. The immune genetic algorithm is treated as an optimization technique to find the optimal solution. The presented algorithm recommends the use of adaptive probabilities of crossover, mutation and immune operation. Furthermore, it effectively exploits some prior knowledge of pending problem and the information of evolved individual's past history to make vaccines. The segmentation algorithm based on the AIGA is implemented and tested on several gray-scale images. The satisfactory experimental results are obtained. In addition, we compare this method with the other segmentation techniques, such as the Otsu's histogram thresholding and the fuzzy c-means clustering. AIGA is found to outperform these two methods.