Study Of Chinese Character Signature Authentication Based The Multi-Agent System

This paper arms to get an offline-authentication method of Chinese underwrite that integrates many programs based the multi-agent system. Some of the electronic signature authentication in being stresses one feature, and absences the study of the manipulation multi-feature. Our study induce multi-agent system, to establish a new way of multi-feature each other harmonizes, and shoot the obvious advance of the speed of the authentication. This is a test of wholesome to the Chinese underwrite authentication.So we selection some representative characters features for analyses. To establish a multi-agent system with every task being analyzed corresponding to certain character features. The final confirmation of the valid character features depends on the co-ordination of the whole system. Another task is to explain and confirm the final character features. Therefore, some signatures are collected, the pretreatment of the signature image is discussed and some statistical features of the Chinese characters are analysized. The features include features of stroke density, features of stroke directions, features of stroke density against the background of the characters, features of the outline of character background, features of the around difference and so on, these features contain the structure information on Chinese characters, so they can be used for analysis as statistical information. We also construct a feature of gray level that can reflect the degree of stroke. Combination the feature of gray level and other features can advance the efficient of electronic signature authentication.Thanks to the maturity of the multi-agent system technology, the electronic signature authentication system based on multi-agent has already become an efficient means to solve electronic signature authentication, and it can solve the problem of the shortage of some electronic signature authentication system which emphasizes the specific feature matching. Making use of the harmony, distributing disposal and simultaneity of the multi-agent system, our system adopts multi-feature matching in order to solve the electronic feature authentic problem more efficiently. Meanwhile, the paper also puts forward a new algorithm for task assignment on signature authentication of Chinese characters based on multi-agent integrating genetic algorithm and stimulated annealing. Furthermore by synthesis of all of the above achievements, a simple experimental system is designed and some signatures are tested.To a certain extent, our work has showed that the signatures that have been so far collected are all successful. Thus it is feasible to solve the electronic signatures authentication problem by making use of multi-agent. However, the popularization of this system needs large-scale testing. From a long-term point of view, this system needs not only large-scale testing but also more features of Chinese character in order to improve the authentication efficiency.