Image quality analysis of electrophotographic printers for banding measurement and forensic application
Abstract
Electrophotographic (EP) process is widely used in the printing industry. One of the major image quality issues in EP printers is known as banding which is related to the fluctuations of the internal printer components. In this thesis, we have proposed a robust and accurate system for banding measurement from the output of an EP printer. The developed tool will measure banding frequencies of a monochrome EP printer from different independent sources. The complete system includes a measurement unit, a specially designed test page, and analysis software. The banding measurement tool will be helpful in the development of new monochrome EP printers by providing a quantitative assessment of banding characteristics for these engines. The banding artifact can also be treated as an intrinsic signature of a printer that can be used to identify a specific printer. Printer identification based on a printed document has many desirable forensic applications. We propose to use the banding measurement frame work and advanced pattern recognition techniques to detect the intrinsic signatures. The projection data from individual characters of a text only document are represented in a reduced dimension by PCA (principal component analysis). Then a classifier is built based on Gaussian mixture model (GMM). The distributions of all the printers help us to determine the mixing coefficient for the projection from an unknown printer. Finally, the decision making algorithm identifies the correct printer.
Degree
Ph.D.
Advisors
Allebach, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.