Abstract
Printer identication based on a printed document has
many desirable forensic applications. In the electropho-
tographic process (EP) quasiperiodic banding artifacts can
be used as an effective intrinsic signature. However, in text
only document analysis, the absence of large midtone ar-
eas makes it difcult to capture suitable signals for banding
detection. Frequency domain analysis based on the pro-
jection signals of individual characters does not provide
enough resolution for proper printer identication. Ad-
vanced pattern recognition techniques and knowledge about
the print mechanism can help us to device an appropriate
method to detect these signatures. We can get reliable in-
trinsic signatures from multiple projections to build a clas-
sier to identify the printer. Projections from individual
characters can be viewed as a high dimensional data set.
In order to create a highly effective pattern recognition
tool, this high dimensional projection data has to be repre-
sented in a low dimensional space. The dimension reduc-
tion can be performed by some well known pattern recog-
nition techniques. Then a classier can be built based on
the reduced dimension data set. A popular choice is the
Gaussian Mixture Model where each printer can be rep-
resented by a Gaussian distribution. The distributions of
all the printers help us to determine the mixing coefcient
for the projection from an unknown printer. Finally, the
decision making algorithm can vote for the correct printer.
In this paper we will describe different classication algo-
rithms to identify an unknown printer. We will present the
experiments based on several different EP printers in our
printer bank. The classication results based on different
classiers will be compared .
