| Originated in the Euphrates and Tigris rivers, in modern Iraq and North Syria as the center of the Ancient Mesopotamia West Asia is the birthplace of human civilization, tens of thousands of cuneiform literature that are unearthed from the region is aware of the world's largest number, at the earliest records of human activities text materials. The successfully deciphered of the mysterious characters which compose by cuneus, marks the establishment of the new study subjects - the Assyrian research by unearthed documents of the Mesopotamia civilization. A large number of original documents provide the knowledge that the predecessors have not extremely rich first-hand information for the Assyrian scholar, but due to the prevailing historical conditions, such as writing material constraints, as well as the handwriting vary from person to person, resulting in a standard text not reunification, and after the passage of time, making some deficiencies found in the literature, the study of the Assyrian scholars pose a great challenge. The use of radial basis function (RBF) artificial neural network on the word recognization, standardization in the literature, its high fault-tolerance shortcomings of the letter can be good repair, the definition of a specific fuzzy model, identify the law of the word's sequence, to the ancient civilization research scholars to create good conditions.In the opening of this paper introduces the Assyriology, the first half describe in detail the origin and development process of the cuneiform, particularly the special cuneus in the evolution of the cuneiform and the basic component (the basic radical radicals compose of the character). Continuation of these elements, the paper introduces RBF artificial neural network, a detailed discussion of the RBF artificial neural network involve in the important subjects. These subjects include the network topology of RBF artificial neural network, the artificial neural by the using, network training algorithms, the training process, and so on.In this next part of the paper highlighted Sumerian cuneiform picture deal with the fuzzy model: All of the basic plans extracted from the Sumerian cuneiform, selected to meet the requirements of specific plans in the basic definition of their point of view and scaling, and other attributes, fuzzy processing. Fuzzy attributes don't exclude by the definition of the basic map of the original Sumerian cuneiform graphics selected match point and rectangular area, when match to achieve the desired accuracy requirements, using the basic map exclude the attributes and fuzzy definition to replace the original character graphics corresponding rectangular area.After the detailed briefings on the fuzzy model, this paper focus on the algorithm model which use RBF artificial neural network to character recognization on the Sumerian cuneiform, RBF artificial neural network use the training sample by fuzzy model. In the use of this model to recognize the standard Sumerian cuneiform and the character in the inscriptions by ancient West Asia Sumer people described on the stone tablet has achieved considerable recognization ratio, as using the model to the practical applications provide demonstration basis; the word's sequence in the matching algorithm process also has an important value for today's highly information of the times. |
