Joel Zlotnick, Jordan Brough and Troy Eberhardt

In Part 1 of this series, the authors described why colour and other artwork cues can help counterfeiters to understand and replicate genuine document artwork.[1] This Part 2 paper explores security artwork design options that deny colour and other cues as points of reference for counterfeiters, with the goal of making plate image separation and artwork replication as hard as possible. It describes security artwork strategies that use overlapping wimages to produce variable density in graphics without widening lines or enlarging graphics, which helps obscure the press setup used to print the original genuine document.

The quick scan-and-print workflow of digital counter­feiting allows more casual counterfeiters to circumvent the detailed duplication of document artwork. However, traditional counterfeiters that aim to replicate the artwork on each individual plate must first determine the number of plates used to print a genuine document. This pre-press step can be made more or less difficult depending on how genuine document artwork is designed. Conventional security design strategies, such as guilloche patterns and security halftones are complex images that require specific design capabilities to generate, while strategies such as split fountains and transparent register require physical offset press capabilities that counterfeiters may not possess. In contrast, the design strategies presented in this paper interrupt counterfeiting pre-press processes in a different way by obfuscating the press setup used to print a genuine document. The goal is to use artwork as a tool to discourage traditional counterfeiting entirely, or at least discourage actual artwork replication in favour of simulation with CMYK or other halftone processes, since these reduce counterfeit quality and can improve the chance of detection.

In Part 1 of this series, the authors analysed examples of security artwork from the point of view of a counter­feiter and identified several characteristics of conven­tional security document artwork that can facilitate artwork reverse engineering, which are described in Table 1a. The authors proposed these cues be mini­mised by adoption of the strategies shown in Table 1b, which deny counterfeiters ready access to information required to separate and replicate individual plate images. In this part of the series, we focus on how overlapping monochromatic artwork can deny colour as a cue for artwork reverse engineering, and how a security design strategy can comply with the recommendations shown in Table 1b.

 

Overlapping monochromatic artwork
The strategy for the prevention of artwork reverse engineering presented in this paper relies on printing layers of artwork from multiple offset plates in a single colour of ink, which is demonstrated using the hexagon graphics shown in Figures 1 through 4. The translucent ink in Figures 1 through 4 is formulated so that up to four sequential layers of this pale ink produce a visibly darker monochromatic image, just as each translucent ink in a CMYK process increases image density as additional colours are printed. Figure 1 shows four unique hex designs, each of which contains a unique interior pattern, and four more hexes that are rotated versions of the first four. Each hex covers approximately the same surface area of the substrate, regardless of the details of its design. If two of the hexes shown in Figure 1 are printed on top of one another from two different offset plates, some possible image combina­tions are shown in Figure 2. Similarly, Figures 3 and 4 show how three or even four of the hex designs from Figure 1 could be printed over one another to create even more complex patterns. Areas where the patterns overlap become incrementally darker, and the hexes in Figure 4 can show as many as four different layers of ink overlap. Additionally, the graphics shown in Figures 2 through 4 are just a sample of a much larger number of possible combinations, which are not all displayed due to space limitations.Figure 1: Template hex designs, each printed in a translucent green ink. Hexes in the bottom row are rotated versions of hexes in the top row. There are two density levels: one for ink and one for blank paper.Figure 1: Template hex designs, each printed in a translucent green ink. Hexes in the bottom row are rotated versions of hexes in the top row. There are two density levels: one for ink and one for blank paper.

Figure 2: Combination artwork formed by printing two of the hexes from Figure 1 on top of one another. The two hexes at the right are bolder because they are composed of the same artwork printed twice. There are three possible density levels: darker green, lighter green, and paper.
Figure 3: Combination artwork formed by printing three of the hexes from Figure 1 on top of one another. There are four possible density levels: three levels of green ink, and blank paper.
Figure 4: Combination artwork formed by printing four of the hexes from Figure 1 on top of one another. There are five possible density levels: four densities of green ink, and blank paper.

Look carefully at Figure 4, and try to determine how many separate plate images are present, and the specific graphics on each plate. Because colour does not provide guidance, reverse engineering Figure 4 is a challenging problem. If the number of plates cannot be determined, then the artwork on each individual plate cannot be identified, and each unique plate design cannot be replicated. Of course, counterfeiters are left with some alternatives for simulation that are described below, including discussion of why the results are likely to be subpar.

Importantly, there is no reason that the units in this example must be hexes, or of uniform shape, or block shapes as opposed to line patterns, or of limited graphical complexity, or of the same size as one another, or of any particular size in print, or even of consistent size across the design. The hex examples in Figures 1 through 4 are a considerable simplification of what is possible with such a monochromatic strategy, which could employ significantly more complex artwork, an unlimited number of unique graphic elements, a greater number of overlapping images and even more complex use of monochromatic colour (which is a topic for future work). The next section will expand upon this hex example to a larger artwork concept that addresses more of the recommendations shown in Table 1b.

Extrapolating to larger designs
Following the hex example above that used as many as four overlapping monochromatic designs to obfuscate the press setup and plate artwork, consider how the recommendations presented in Table 1b can be imple­mented in a larger and more complex design based on the hex theme. Figure 5 shows the artwork on each of the four printing plates in a hypothetical four-station offset press, and Figures 6 and 7 show the composite artwork that results when all four plates are printed together.

Figure 5: Four individual printing plate images, which are added together to form the composite image in Figures 6 and 7. Each plate image uses the same graphic style to prevent any of them from standing out.

Figure 6: Composite artwork from overprinting of the four plates shown in Figure 5. Areas of higher density are created not just by covering more surface area, but also by overlapping multiple layers of green ink. 
Figure 7: Magnified view of part of the composite image shown in Figure 6, illustrating how darker areas are created by over­printing images from different plates

As in the hex example, the same translucent green ink is printed from each of the four fountain units, so a counterfeiter cannot differentiate between plates on the basis of colour. Further, the designs do not contain large continuous traceable line patterns, since the hexes are separated by white space, so understanding the artwork of one hex does not necessarily help with understanding the artwork of other hexes. Additionally, many of the hex designs in the composite artwork in Figure 6 are composed of two, three or four plate images overlapped with one another, though some are printed from only one plate. (And, certainly, there is a case that every single hex in Figure 6 should be printed by at least two plates, though we have not done so in this example.) Finally, the same basic pool of hex designs is used on all of the four plates shown in Figure 5, so each plate does not contain graphical cues that help distinguish one plate from another plate.

Consider how image density is created in Figure 6. Figure 6 contains areas of low and high density even though the surface area covered by at least one layer of artwork is consistent through most of the design, as is the spacing between the hexes. In contrast, conven­tional security printing strategies use a single layer of ink per colour, and image density is increased by covering more of the substrate with ink (by widening lines, for example). The magnified view in Figure 7 shows how areas of higher density in this example are produced not only by covering more substrate with ink, but also by layering ink to darken the print. Again, how difficult is it to determine how many plates are present in Figures 6 and 7, and the specific artwork used on each?

Counterfeiting and simulation
The monochromatic design example described above requires a more specific explanation of how it disrupts counterfeiting workflows. Although simulation will always be possible, deliberately designed genuine artwork can work to limit simulation quality. This monochromatic design strategy forces counterfeiters to choose between achieving accurate artwork detail and accurate tonal reproduction in simulations, if they fail to accurately replicate the plate artwork. First, the composite artwork in Figure 6 denies counterfeiters access to information that facilitates high-quality art­work replication, including the use of colour, line tracing and artwork style as cues to distinguish between artwork from different plates. If a determined counter­feiter cannot isolate and accurately replicate the individual plate artwork, options for artwork simulation include:

  • simulating the composite line art design with a dot halftone;
  • printing the entire monochromatic design from a single plate (instead of four plates);
  • reverse engineering only the density levels in the final printed artwork instead of the actual plate images.

Simulating security artwork with a halftone process converts the design into one or more patterns of dots that vary in size (amplitude-modulated halftones) or spacing (frequency-modulated halftones). Halftones make it possible to preserve the general tonal values of an image at a macro level, while causing loss of image detail at the microscopic level. For example, Figure 8 shows the artwork from Figure 6 as it could appear if simulated with an AM halftone process.

Figure 8: Simulation of the artwork shown in Figure 6 with an amplitude-modulated half­tone process. This approach preserves more of the overall tonal values at the cost of artwork fidelity. Compare to Figures 6 and 9.

Figure 6 required four printing plates to display the various density levels, but in Figure 8 the counterfeiter has collapsed the artwork onto a single plate. Although the overall density values are preserved by the halftone, the line art is not, and the dots are particularly visible in the lightest areas of Figure 8. Detection of the counterfeit in Figure 8 is easy with a magnifier, since the detail in the artwork is compromised.

In the second option, a counterfeiter could redraw the line art and print it from a single plate, because the design contains only one colour. While the fidelity of the line art in this example can be preserved to a greater extent than with a halftone process, a single plate is not capable of replicating the several density values shown in Figures 6 and 7. A single-plate counter­feit of the image in Figure 6 is shown in Figure 9, which demonstrates how some of the microscopic artwork detail has been preserved, but the tonal variation has been lost, causing Figure 9 to appear posterised and flat compared to Figure 6. The halftone process shown in Figure 8 preserves more of the tonal variation at the cost of artwork fidelity, while the line art strategy shown in Figure 9 preserves more of the line artwork at the cost of the tonal variation.

Figure 9: Simulation of the artwork shown in Figure 6 with line art printed from a single plate. This approach preserves some detail in the line art, but loses much of the tonal value.  Compare to Figures 6 and 8.

The third counterfeiting option is to simulate the density levels present in the genuine image. The counterfeiter creates a single plate for each discrete density level and prints using multiple inks of the same colour but different densities. A simple example of such a counter­feit executed on only two plates (instead of four) is shown in Figures 10 through 12, where Figure 10 prints a light green ink, Figure 11 prints a dark green ink and Figure 12 shows the two plates printed together in register. Although the approach shown in Figure 12 offers better simulation of both artwork and density values than the simulations in Figures 8 and 9, the process is more technically complex and can introduce registration problems if four density levels are simulated on four plates, instead of just two density levels as shown in Figures 10 through 12. The authors regard registration as an area of proficiency for legitimate security printers, and genuine document artwork could also be designed to be more tolerant of minor registration variations than the artwork shown in Figures 6 and 7. However, counterfeiters attempting to simulate density (instead of the real plate artwork) risk the appearance of gaps between plate images if registration is not perfect. Figure 13 shows the same two-plate image as in Figure 12, but with the plates slightly out of register.

Figure 10: Part of a density simulation of the artwork shown in Figure 6. This artwork would be printed in light green ink and paired with the artwork shown in Figure 11 to create Figure 12.
Figure 11: Part of a density simulation of the artwork shown in Figure 6. This artwork would be printed in dark green ink and paired with the artwork shown in Figure 10 to create Figure 12.
Figure 12: A two-plate density simulation of the artwork shown in Figure 6. This artwork contains two designs printed from two plates: a light ink from Figure 10 and a dark ink from Figure 11. Compare to Figure 6.
Figure 13: A two-plate density simulation of the artwork shown in Figure 6, with the light and dark plate images printed out of register, resulting in gaps where the plate images meet. Compare to Figures 10 through 12.

A density simulation attack more sophisticated than the one shown in Figures 10 through 12 is possible if a counterfeiter can identify how many discrete density levels are present in the image, which is only possible in Figures 6 and 7 because there are just four plates in this simple example. The authors have already iden­tified solutions to this risk that exceed the scope of this paper, but in summary these involve increasing the number of discrete density levels available where the plate images overlap. This can be accomplished by delivering inks of the same colour but different trans­lucencies from the four fountains, and by utilising split fountains for transitioning ink translucency instead of colour. Both of these strategies will be discussed in greater detail in future work.

Conclusion
Adoption of a monochromatic design strategy drastically restricts colour gamut and imposes significant creative constraints, but also provides many security benefits. The monochromatic design example described in this paper actively resists attempts to reverse engineer its plate artwork, and reproduces poorly when subjected to typical traditional counterfeiting and simulation strate­gies. In the area of design, a monochromatic artwork strategy offers designers different techniques for control of image density, which can be varied by changing line width and also by printing multiple layers of the same ink, providing new opportunities for exquisitely detailed monochromatic designs. Although this paper has focused on options that exist within the considerable constraint of a single translucent ink formulation, future work will consider variable ink translucency, split fountains for transitioning density instead of colour, and revisiting of colour as an additional variable in an overarching strategy to make security artwork increasingly resistant to reverse engineering.

MORE ABOUT THE AUTHORS

Joel Zlotnick is employed by the US Department of State, Bureau of Consular Affairs, Counterfeit Deterrence Laboratory as a physical scientist. He conducts research on how design strategies can help maximise the security value of printing technologies and security features, and develops training programmes on counterfeit detection.

Jordan Brough is employed by the Homeland Security Investigations Forensic Laboratory as a forensic document examiner, specialising in adversarial analysis and counterfeit deterrence. Jordan spends his time examining suspect documents and consulting with United States security document designers.

Troy Eberhardt is employed by the US Immigration & Customs Enforcement Homeland Security Investigations Forensic Laboratory. He supervises the Research and Development Section at the laboratory, which specialises in identifying and mitigating vulnerabilities within travel documents.

Reference

1 Zlotnick, J., Brough, J. and Eberhardt, T. (2016). Interrupting Traditional Counterfeiting Workflow, Part 1: Colour and Split Fountains.