signatures, much less attention has been paid to o!-line
processing. The on-line system produces time information
like acceleration (speed of writing), retouching, pressure
and pen movement. It already has recognition and
veri"cation rate of 100%. Therefore, nothing of value can
be added in this "eld. On the other hand, most of this
information is lost in the o!-line system. However, other
useful factors which can be used to di!erentiate the
handwriting of one person from another still exist.
O!-line signature processing remains important since
it is required in o$ce automation (OA) systems. It is used
for the validation of cheques, credit cards, contracts,
historical documents, etc. Since the signature is processed
as an image, there is no great di!erence between Arabic
signatures and other signatures but experiments have
shown that some features are not e!ective in Arabic
signature recognition and veri"cation.
The purpose of the recognition process is to identify
the writer of a given sample, while the purpose of the
veri"cation process is to con"rm or reject the sample.
Machine recognition and veri"cation of signatures is a
very special and di$cult problem. The di$culty arises
from the following [3]: