Research Of Off-line Signature Verification System By Computer

During the off-line signature appraisal , few dynamic information can be available , the difficulty of the identification is tremendous. At present, the development of this field is slow. Due to a large number of cases which are related to the signature , a automatic identification system is needed badly.The system which is mentioned in this thesis is based on the software VC ++6.0 which act as the fundamental programming environment , having realized some basic functional modules and some key technologies and functions.Especially, the function of filtering background in the course of original signature image pre- processing can extract the colorful signature handwriting promptly. It also has the function of extracting the handwriting which have different color under the complicated signature background. Its main features are high information content, little distorted, convenient etc.During the image processing of acquiring two-value signature image, traditional methods aren't adopted, the colorful signature handwriting extracted is processed directly into two-value image. The system's processing efficiency and effect can be improved.In false dynamic characteristic , the function of extracting signature high grey level is realized by program. The slope variety of the width distribution is put forward as one of false dynamic characteristics which can reflect the dynamic information such as velocity variety etc effectively. Due to the feature is added ,the ability of the system identification is improved rapidly.Through the combination of form and false dynamic characteristics , the rate of false recognition can be reduced effectively , preferable comprehensive identification effect can be achieved.