Paradigm shift in document related frauds: Characteristics identification for development of a non-destructive automated system for printed documents
Introduction
“One can not come into contact with an environment without changing it in some way.”
–Exchange Principle, now called Contact Traces, first articulated by Edmond Locard in 1910.
This is still an age of hardcopy documents, and a paper-less world is still a distant dream. We depend on these printed documents in many of our encounters with the complexities of modern life. Hardly a day goes by without some document playing a crucial part in the life of every one of us. Therefore, it is of no surprise that criminals expend significant effort and resources to create fraudulent documents.
Document manipulation and tampering have seen a paradigm shift in the form of Digitized Document Frauds (DDF). Digitized Document Fraud is defined as the process of scanning any conventional document, editing the scanned image to get the desired changes using image processing software and finally printing the document using printer. Examples of DDF include generating fake currency, stamp papers, and other documents like educational certificates, degrees, cheques, will, property papers, licenses, immigration and visa documents, security passes and badges. One of the famous and recent frauds is the billion dollar fake stamp paper scandal, which rocked India in 2004; this highlights the burning problem of misuse of digital technology in Conventional Document Frauds. The other famous cases include generation of fake postal stamps, which are generally used only once, thus costing heavily to the exchequer, fake transcripts and experience certificate generation for availing of over privileges in educational and private institutes to name a few.
The use of sophisticated computerized systems capable of producing high quality hardcopies makes the job of criminals much easier and safer than the conventional ways of tracing, imitating and tampering of documents due to the former being extremely difficult to detect using conventional mechanisms, as the output is almost the same as the real document. Because the fraudulently generated documents are so realistic, there are less chances of the forger being caught, making Digital Document Fraud a low-risk high-gain venture.
The objective of this study is to detect and attribute the document to the machine and tool used for its generation. Until now there were no well-defined procedures and guidelines to forensically link the crime to the perpetrator in Digitized Document Fraud cases. Our work is based on the following principles (Chisum and Turvey, 2000, Chisum, 1999, DeForest et al., 1983, Saferstein, 1998, Inman and Rudin, 2001):
- 1.
Locard's Exchange Principle of Evidence, which states “One can not come into contact with an environment without changing it in some way”.
- 2.
Individuality principle, which states that no two objects are identical. It is often the case that two objects cannot be told apart, but they are not identical. If the two objects are distinguishable, it is obvious that they are not from the same source. However, if they are indistinguishable, they need to be examined in more detail to determine whether they are from the same source.
- 3.
Every effect has a cause. The cause must precede the effect.
- 4.
Principle of comparison is based on the theory of comparison of questioned information with the genuine (original) one and with suspects (if available) using existing tools and technology. It states that we can almost always detect and fix the culprit if enough of all genuine, questioned and suspected material is available; also the innocence can be proved, in case of false implications.
In the light of above-mentioned principles this study tries to identify the characteristics unique to a scanner/printer. This analysis of characteristics will lead to linking of crime to the criminal. The identified peculiar characteristics can serve as investigatory leads to determine the origin of a document and thus help ascertain the authorship of the documents under question.
This paper explores the role that scanners/printers play in the generation of fraudulent documents. The identified characteristics are specific to printers and scanners and differ considerably even within the same make and same model number. These characteristics are universal and may help in addressing the following issues about the document in question:
- 1.
Is the document in question genuine or fake?
- 2.
Which type of scanner/printer has been used to generate a fraudulent document?
- 3.
Providing clues to locate the suspected scanners and printers.
This forms the basis for the development of a non-destructive automated system for efficiently detecting and fixing the origin and thus authorship of questioned documents. Our findings are effective even in the scenario where an intelligent criminal destroys digital evidence from the storage media of computer system and only the fake document is available to establish and link the crime to the criminal. Provided the proposed methodology can be generalized, its efficiency and reliability of results could help digital investigators combat this growing threat to economy and society.
Section snippets
Previous work
Detection of tampering in digital images could be useful when the remnant-scanned image is available in storage system. The Popescu, 2004, Luxen and Forstener, Fridrich et al. (2005), Farid, 2002, Popescu and Farid, in press-a, Fridrich, Farid and Lyu, 2003, Popescu and Farid, 2005, Popescu and Farid, in press-b could provide vital information regarding the malicious editing in the scanned image. These could help in pinpointing the details of the changes in the fraudulently generated document.
Analytical devices
For the purpose of pattern identification, we used Video Spectral Comparator (VSC), and high-resolution LEICA MZ8 and LEICA MZ 12.5 microscopes (for examining fine dots of various colors), which have a mountable analog camera as well as direct input to digital computer.
The VSC has limited magnifying power, hence the LEICA MZ8 and LEICA MZ 12.5 (Tricolor Zoom Stereo Microscope with DC 300F camera) were used to capture magnified image of a single character from all the documents. The LEICA
Terminology and observations
P1 stands for Printer HP840C used in document generation
S1 stands for Scanner HP300C used in document generation
P2 stands for Printer HP930C used in document generation
S2 stands for Scanner HP6300C used in document generation
This terminology is used for Fig. 1 and Annexure I.
Detailed explanation
The color of pixels depends on the scanning/printing device used; in case of the original document (where scanner has not been used), the color of pixels varies within a very short range of RGB, whereas, in case of a fraudulently generated document (using a scanner and printer), the variation is much more pronounced. These dots originate mostly during the process of scanning/printing and form the basis of differentiating between the fraudulent document and the genuine. The pattern of dots in
Conclusion
The identified characteristics could be used to develop a non-destructive automated system for tackling the paradigm shift in document related frauds. The proposed automated system using the above-identified characteristics can take the originals (admitted and specimens) into consideration along with the questioned documents as the input to generate efficient results. This would be done by processing and comparing, and then finally segregating the documents into separate categories. Hence we
Future work
Present study uses combination of two scanners and inkjet printers of HP to uniquely detect and fix the device used. The other devices of Canon, MODI, and HP (both inkjet, LaserJet and color photocopier) were used to confirm the presence of characteristics universally. Now more scanners and printers of various make and model need to be examined, so that we can come up with a signature database to say about the type/make/model/piece of scanner/printer used in DDF, without availability of the
Acknowledgements
We thank Mr. Anirban Sengupta for his comments and recommendations in preparing the final version of this paper.
We also thank the office of the Government Examiner of Questioned Documents, Hyderabad and Kolkata and Directorate of Forensic Science for their support and equipments.
References (20)
- et al.
Evidence dynamics: Locard's exchange principle & crime reconstruction
Journal of Behavioral Profiling
(January, 2000) Criminal profiling: an introduction to behavioral evidence analysis
(1999)- Davis LS. Image texture analysis techniques – a survey. University of Texas at Austin, Department of Computer Sciences,...
- et al.
Forensic science: an introduction to criminalistics
(1983) - Farid Hany, Lyu Siwei. Higher order wavelet statistics and their application to digital forensics. In: IEEE workshop on...
Detecting digital forgeries using bispectral analysis
(2002)- Foster and Freeman, UK....
- Fridrich Jessica, Soukal David, Lukas Jan. Detection of copy-move forgery in digital images. Binghamton, NY;...
- Fridrich Jiri. Methods for tamper detection in digital images. SUNY Binghamton, NY and MRC;...
- et al.(1992)