Abstract: In this paper, discrimination method for inkjet printed counterfeits and feature extraction method of spur marks are proposed. Spur marks are tool marks formed on a printout by paper conveyance gears of inkjet printers, and inkjet printers are classified into some groups by type of spurs and their arrangement. Spur marks were visualized and clearly distinguished from background by infrared oblique lighting and gradient image processing. As the characteristic values of spur marks, the pitch and mutual distance were introduced. By radon transform of spur mark images, direction of spur mark lines and the mutual distance were estimated. Spatial frequency components of spur marks were analyzed by two-dimensional fast Fourier transform to estimate the pitch. The pitch and mutual distance were extracted almost automatically. Proposed methods were applied to samples printed by a color inkjet printer. Measurement results were compared with those obtained by conventional method using x-y stage, and the accuracy was proved. By detecting existence of spur marks on printouts, discrimination of inkjet printed material was achieved. This technique will be applicable to model and manufacturer identification of printers, and significant not only for counterfeit detection but also in the field of forensic document examination.

@conference{akao2002discrimination,
  author       = {Y. Akao and K. Kobayashi and S. Sugawara and Y. Seki},
  url          = {http://spiedl.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=PSISDG004677000001000129000001&idtype=cvips&gifs=yes&ref=no},
  booktitle    = {Proceedings of SPIE},
  volume       = {4677},
  year         = {2002},
  title        = {{Discrimination of inkjet-printed counterfeits by spur marks and feature extraction by spatial frequency analysis}},
  pages        = {129},
}