Study On The Artificial Immune System Applied To The Hand-written Number Recognition

The reaearch woke in this paper not only presents a new thought and method for the techne of hand-written number recognition, which has high theoretical value and practical significance, but also explorates and attempt to the further research and application of artificial immune system which is just starting but has a great application foreground in the image recognition and pattern recognition.In this paper, the application of artificial immune system in the hand-written number recognition is researched, based on the learning about the domestic and overseas present research and the extensive collection of prevenient research data, the immune mechanism and algorithm is lead in the hand-written number recognition, and all of the elements of immune system is defined, then the relative description of the problem of the hand-written number recognition is corresponded with the relative concept of immune system and the immune principle what is used in the system, and the mathematics expression of relative immune mechanism is defined, after that, the artificial immune system model used for the problem of the hand-written number recognition is designed, the model has features of study, memory,self adapting and diversity, and the corresponding algorithm of the model is gived. The simulation experiment on number samples after pre-processing and feature extraction of the self collective hand-written number sample set is done with Matlab6.1, and the result shows that the artificial immune system model used for the problem of the hand-written number recognition posed in this paper is capable of recognition to certain extent, and the corresponding algorithm is operated easily. If the relative variables can be choosed properly, better recognition effect of the system can be get. At last, the next research wok is previewed.