| Summary: | 碩士 === 國立成功大學 === 資訊工程研究所 === 81 === Signature verification is a professional skill in tradition,
the work must rely on an expert and this has constrained its
application. Because computers and their applications have make
many complex things more efficient and easier in the following
years, we also tried to apply computer techniques to construct
an On-line Chinese Signuture Verification System. From this
research, we can explore the possibility of substitute computer'
s techniques for traditional signature verification skills, and
expand the computer's applications. This On-line Chinese
Signature Verification System is based upon the Bayesian
network. The structure and performance of Bayesian network are
very similar to a traditional neural network. It must be
trained before being used. But the training process of Bayesian
network is very fast, while it may take a very long time to
train a traditional neural network. In this thesis, we also
developed some techniques to extract the features of a
signature , including dynamic and static ones. And from the
experimental results, we found they are effective.
|
|---|
