Handwriting is a traditional method of personal authentication. In China, people write documents adopting their favorite calligraphy, and thus their handwriting is permeated by their emotion and style.

       From the perspective of a researcher, handwriting is the synthetic reflection of personal physiological attributes together with his postnatal learning process. This process is unique to everyone, and researchers are trying to grasp these individualities as a method of personal identification.

       The research on handwriting is prompted by the development of new handwriting acquisition devices. Just as the adoption of scanners of high resolution boosted off-line handwriting analysis years ago, digital tablets with high sampling accuracy have switched the focus of handwriting research to the on-line handwriting nowadays. This trend has led to many interesting products, among which digital signature has already been introduced to practical utilities.

       Current research on handwriting identification tackles two families of problems, namely text-dependent handwriting analysis and the text-independent case. When considering the text-dependent handwriting, we often resort to the writing content for reference, which may be of great help to the analysis, but we must keep in mind that the indicated content of handwriting has restricted the utilities of pertinent methods. In the text-independent case, the writer is allowed to write at will with no constraints on the writer, and this is the instance of our practical writing. But without the knowledge of writing content, the problem became much more complicated. And that's the main reason why no practical products for text-independent handwriting analysis have been launched in market. Consequently, our research is focused on this promising area.

       Current work on text-independent handwriting analysis is not progressing as the expectation of a common user, because from a researcher's point of view, there are not many methods for solving the identification problem. But if we could find effective strategies for text-independent handwriting analysis, or improve on the existing methods, for example, texture analysis and fractal analysis, the prospective for application is inspiring. Products derived from the methods are in great demand in areas such as judicature, finance and police. Moreover, text-independent handwriting bears more advantages than the text-dependent case, including the depressed demand for database secrecy, and increased difficulty for simulation.

       Our proposed method for this problem extracts the dynamic features from the record of the writing sequence, where pressure, position, and writing velocity etc. are considered. The processing of the handwriting is arranged stepwise on the stroke level. The matching procedure is implemented with the reliability of the respective features considered. Presently, the temporal varieties of writing styles and the influence of writing circumstance on the writer attract our research interest, where we hope to pinpoint the essence of writing.

       Because of the different cultural background of writing, western researchers put more emphasis on western document analysis instead of oriental cases, and the research groups focus on oriental characters are mainly in Asia. With the most population in the world and without mental resistance to handwriting, China bears the best potential market value for handwriting products, which lays a solid foundation for our research on automatic analysis and decision of the validation of handwriting. And such automation would not only detect attempts at fraud in handwriting, but also but also greatly discourage such attempts.