Survey of Scanner and Printer Forensics at Purdue University

Abstract

Abstract. This paper describes methods for forensic characterization of scanners and printers. This is important in verifying the trust and authenticity of data and the device that created it. An overview of current forensic methods, along with current improvements of these methods is presented. Near-perfect identification of source scanner and printer is shown to be possible using these techniques.

Accurate and Robust Identifying Forged Region Method in Scanned Images

“…It is clear from the table that the highest detection accuracy for large and small forged regions is obtained in the proposed method; however the block size used in the proposed method lies between the other two published methods. The detection accuracy for [11] is higher than that of the method used in [14] for large forged regions but its accuracy is smaller than that of [14] for small forged ones. The FA of the proposed method lies between the two other methods, but it has been detected with the highest and small false detection.…”

“…The first proposed method used for forged region detection is [11] and uses the statistical feature of sensor pattern noise to identify the forged regions on only one group of tampered images. This group is similar to our first group.…”

“…This group is similar to our first group. Figure 9 shows one of the tested examples in [11]. There are some errors in missing some forged regions or detecting original regions as forged ones.…”

“…In [7][8][9][10] support vector machine (SVM) has been trained with these features to identify the source scanner. In [11] SVM has been trained again by these statistical features to detect forged regions in the scanned images. The forged images had been synthesized using scanned images from different scanners.…”

Text-line examination for document forgery detection

“…Several uses of intrinsic features have been presented in previous publications: printer identification, the process of assigning a printout to a unique printer or a printer type, has been intensively studied by different groups. Mikkilineni et al [8][9][10] present a set of gray scale features that is used to determine the type of laser printer that was used to print a document. Schreyer et al [11][12][13] worked on detecting the printing technique used to print a document.…”