Immune Programming And Its Research On Applications In Image Segmentation

Artificial immune system (AIS) is a novel branch of computational intelligence research area, and gradually becomes a rising hot point of evolutionary computation and pattern recognition in recent years. Because of the features of particular self-learning, self-organizing and high parallel mechanisms, it is natural to view AIS as a powerful information processing and problem -solving paradigm, which paves a new way to solve the hotspots and difficult problems in the engineering fields.Immune programming (IP) is a novel idea of utilizing AIS in engineering application, which is combined of immune mechanism and evolution mechanism. It can construct immune operator by utilizing local characteristic information to intervene the parallel global search with certain intensity. Therefore, the designed algorithm shows wonderful global convergence capability. Besides, IP has restrained the degenerative phenomena in existing evolution algorithms and greatly improved the algorithm efficiency.Begin with reading a lot of documents and thinking deeply, the research work of this thesis emphasized particularly on the feasibility of applying IP in the image segmentation. By designing the image thresholding segmentation algorithm based on the maximum entropy theory and the image segmentation based on minimizing of the expense function, IP is applied successfully in the image segmentation. Not only does it fit for images with double-peak-shaped histogram, but also those of complicated-shaped histogram. Theoretical analysis and experimental results show that immune programming outperforms the existing optimization algorithms in alleviating the degeneration phenomenon. It is a feasible and effective image segmentation algorithm.