A Discrimination Method of Paper by Fourier Transform and Cross-Correlation (Part 2)
Makoto Shinozaki, Hitomi Miyata, Tomohito Nakayama, Toshiharu Enomae · JAPAN TAPPI JOURNAL · 2001
The previous report by some of the authors implied a possible usefulness of the image analysis technique for quantitative discrimination of paper under different brand names. Image analysis is a none-destructive discrimination technique and desirable for forensic purposes. Our previous work employed a cross correlation matching method for the power spectrum of 2-dimensional fast Fourier transform on light transmitted images of paper sheets. This report presents an application of that method to the discrimination of commercial PPC paper under different brand names. PPC paper, i.e. plain paper copier paper or xerography paper, is being produced increasingly and consumed for more variety of purposes these days. PPC paper has been used mainly for black-and-white photocopy in early years but now it is used also for color photocopy, for personal and/or SOHO laser printers and for inkjet printers.Commercial PPC paper under 12 different brand names provided by different manufacturers were examined as specimens in this work. Ten of the twelve sample and reference files correctly showed the largest similarity when they were correlated with the files obtained under the same brand name. Two pairs of sample and reference files, however, did not show the largest similarity. Both of the pairs showed resembling images when they were inverse Fourier transformed so that typical patterns would be extracted, which means those pairs could have been manufactured from the same machines, respectively.The results showed that this method of image analysis with Fourier transform was able to discriminate commercial PPC paper under different brand names or their manufacturers.