Digital forensics of printed source identification for Chinese characters

Abstract

Recently, digital forensics, which involves the collection and analysis of the origin digital device, has become an important issue. Digital content can play a crucial role in identifying the source device, such as serve as evidence in court. To achieve this goal, we use different texture feature extraction methods such as graylevel co-occurrence matrix (GLCM) and discrete wavelet transform (DWT), to analyze the Chinese printed source in order to find the impact of different output devices. Furthermore, we also explore the optimum feature subset by using feature selection techniques and use support vector machine (SVM) to identify the source model of the documents. The average experimental results attain a 98.64 % identification rate which is significantly superior to the existing known method of GLCM by 1.27 %. The superior testing performance demonstrates that the proposed identification method is very useful for source laser printer identification.

Appendix

The revised/adjusted formulas in this study and in Mikkilineni [15]

	The formulas in this study	The Formulas in Mikkilineni [15]
Revised formula	$GLDHenergy={\displaystyle \sum_{k=0}^N GLDM{(k)}^2}$	$GLDHenergy={\displaystyle \sum_{k=0}^N GLDM(k)}$
	$GLSHenergy={\displaystyle \sum_{k=0}^{2N} GLSH{(k)}^2}$	$GLSHenergy={\displaystyle \sum_{k=0}^{2N} GLSH(k)}$
	$\begin{array}{l} GLSHshade=\\ {}\kern1.5em {\displaystyle \sum_{k=0}^{2N}\frac{{\left(k-{\mu}_x-{\mu}_y\right)}^3 GLSH(k)}{{\left({\sigma}_x^2-{\sigma}_y^2+2\rho {\sigma}_x{\sigma}_y\right)}^{\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}}\end{array}$	$\begin{array}{l} GLSHshade=\\ {}\kern1.5em {\displaystyle \sum_{k=0}^{2N}\frac{{\left(k-{\mu}_x-{\mu}_y\right)}^3 GLSH(k)}{{\left({\sigma}_x^2-{\sigma}_y^2+2\rho {\sigma}_x{\sigma}_y\right)}^{\raisebox{1ex}{$3$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}}\end{array}$
	$\begin{array}{l} GLSHprom=\\ {}\kern1.25em {\displaystyle \sum_{k=0}^{2N}\frac{{\left(k-{\mu}_x-{\mu}_y\right)}^4 GLSH(k)}{{\left({\sigma}_x^2+{\sigma}_y^2+2\rho {\sigma}_x{\sigma}_y\right)}^{\raisebox{1ex}{$4$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}}\end{array}$	$\begin{array}{l} GLSHprom=\\ {}\kern1.25em {\displaystyle \sum_{k=0}^{2N}\frac{{\left(k-{\mu}_x-{\mu}_y\right)}^4 GLSH(k)}{{\left({\sigma}_x^2-{\sigma}_y^2+2\rho {\sigma}_x{\sigma}_y\right)}^{\raisebox{1ex}{$4$}\!\left/ \!\raisebox{-1ex}{$2$}\right.}}}\end{array}$
Adjusted formula	${\mu}_x={\displaystyle \sum_{i=0}^Ni\times {p}_x(i)}$	${\mu}_x={\displaystyle \sum_{i=0}^N{p}_x(i)}$
	${\mu}_y={\displaystyle \sum_{j=0}^Nj\times {p}_y(j)}$	${\mu}_y={\displaystyle \sum_{j=0}^N{p}_y(j)}$
	${\sigma}_x^2={\displaystyle \sum_{i=0}^N{\left(i-{\mu}_x\right)}^2{p}_x(i)}$	${\sigma}_x^2={\displaystyle \sum_{i=0}^N{i}^2\times {p}_x(i)-{\mu}_x{}^2}$
	${\sigma}_y^2={\displaystyle \sum_{j=0}^N{\left(j-{\mu}_y\right)}^2{p}_y(j)}$	${\sigma}_y^2={\displaystyle \sum_{j=0}^N{j}^2\times {p}_y(j)-{\mu}_y{}^2}$