Joel Zlotnick, Jordan Brough and Troy Eberhardt

Reverse engineering of document artwork is a critical prepress activity for sophisticated counterfeiters that print by offset lithography or other non-digital techniques, as described in Part 1 of this series.[1] In Part 2, several techniques were recommended that make artwork reverse engineering more difficult, including delivery of the same ink from multiple printing plates as part of a monochromatic design strategy.[2] As this strategy is limited in the number of discrete density levels, this Part 3 paper explores ink opacity as a variable that can expand the number of density levels available, resulting in increased image complexity and better resistance to reverse engineering.

Earlier work in this series described how traditional counterfeiters study artwork to determine how a genuine security document was printed, because they must understand the press setup before replicating artwork on the individual printing plates. Security designers possess considerable opportunity to conceal the nature of the press setup and plate artwork, which can force even skilled counterfeiters to simulate rather than replicate artwork, resulting in inferior counterfeits.

Part 2 of this series presented a monochromatic security design strategy that targets four visual cues which counterfeiters use to understand how a genuine document was printed. These include:

  • printing of different ink colours from each plate;
  • use of large continuous line patterns that can be visually traced;
  • delivery of complete visual elements from a single printing plate and;
  • use of different artwork styles on different plates.

The systemic elimination of these four cues from security artwork results in a monochromatic block design featuring extensive overlap between different printing plates, with every plate containing similar graphic elements.

An element of this monochromatic strategy is the delivery of the same translucent ink colour from every fountain, where areas of darker print are created by sequentially printing multiple layers of translucent ink over one another. This is the opposite of typical practice in security printing, where darker areas are made by widening lines or otherwise covering more of the substrate with a single layer of more opaque ink. The monochromatic strategy produces artwork that is hard to understand and difficult to reverse engineer, but only produces a number of discrete ink layer opacities equivalent to the number of printing plates used, since every fountain contains exactly the same ink. This part of the series addresses this limitation by considering ink opacity as a tool to expand the number of discrete density levels available on a press limited to only four fountains.

A real-world graphic arts workflow must consider important production factors that cannot be accounted for in this short conceptual paper, and some caveats are required. For example, we assume that ink vehicles are colourless and transparent, the density and colour of printed images are not influenced by substrate colour, that inks containing different pigment volumes behave similarly on press, that the subject document is compatible with liberal application of multiple ink layers, and so forth. Most importantly for the purposes of this paper, we assume that the combined opacities of overlapping ink layers are always additively linear regardless of the order of printing, that ink layer thickness can be tightly controlled on press and that changing the pigment concentration will change the opacity of an ink without also changing its colour. Despite these oversimplifications, some of which are inaccurate in real print environments, we believe that security printers can adapt these theoretical concepts to real print workflows through colour management and empirical testing.

Figure 1: A set of hex elements with similar (but not identical) artwork

Ink opacity and image density
Printing ink formulations include pigments, vehicles, solvents and a number of other chemical components that control the colour, rheology, drying mode and other properties of the finished ink. The pigment component is most responsible for the ink colour, though pigments may compose only a fraction of a printing ink formulation (approximately 15-20%, as a general example). The translucency of an ink is a function of its formulation, with higher pigment volumes corresponding to greater opacity for a given ink layer thickness, and lower pigment volumes producing an ink of greater translucency. When inks of different pigment concentrations are deployed in each of four stations in a typical offset press setup, each plate prints an image of different opacity.

Figure 2: Security design plate artwork printed with ink of 8% opacity.

Consider a hypothetical maroon ink formulation in which the pigment volume can be adjusted to produce inks of varying opacity, up to a maximum opacity of 100% (again, assuming no change in colour as pigment density is varied, and no contribution to colour from the vehicle). Four maroon inks are prepared that possess 8%, 18%, 30% and 44% of the formulation’s maximum opacity. Further assume that ink opacities are additively linear, such that printing an ink of 8% opacity on top of an ink of 18% opacity results in a combined opacity of precisely 26%, regardless of which ink is printed first (of course, print order does matter in real applications). Following this logic, layering all four inks produces a final image of 8% + 18% + 30% + 44% = 100% density. The example percentages were contrived to sum conveniently to 100%, and may not represent an optimal set of real densities. Real-world ink opacities (and an optimal real-world print order) could be determined empirically.

Figure 3: Security design plate artwork printed with ink of 18% opacity.
Figure 4: Security design plate artwork printed with ink of 30% opacity.

 

Figure 5: Security design plate artwork printed with ink of 44% opacity.

Each of these maroon inks is deployed in one of four stations in a hypothetical offset press, so each plate prints an image of a different opacity. Figure 1 shows how a set of hex elements with similar (but not identical) artwork, each printed using inks of different opacities, combine so the interior of each hex contains several different image densities, up to a maximum opacity determined by which inks are combined. Figures 2 through 5 show four hypothetical document artwork designs based on a crane and hex theme, where each segregated shape is separated from the others by white space (to prevent line tracing) and contains a slightly different combination of artwork and opacities from nearby shapes. Display of darker print in the composite image shown in Figures 6 and 7 relies on applying multiple translucent ink layers to the same substrate area (instead of covering more of the substrate surface with a single ink formulation). Each shape in Figures 6 and 7 contains images from as many as four plates, but not always all four.

Figure 6: Composite artwork from overprinting of the four plates shown in Figures 2 through 5. Areas of higher density are created primarily by over­lapping multiple layers of translucent ink, not by widening lines or covering more of the substrate surface. Starting only from this image, it is very difficult to work backwards and isolate the artwork on each of the original four plates.
Figure 7: Magnified view of the composite image shown in Figure 6, illustrating how inks of different opacities combine to create a complex image containing multiple densities. This image contains as many as fifteen discrete ink density levels.

Figures 6 and 7 contain fifteen discrete densities that depend not only on the number of plate images over­lapped, but also which particular inks have been combined. Figure 8 shows the fifteen possible combi­nations, from each plate alone up to all four plates in combination. Some of the shape elements shown in Figure 6 are printed from only one plate, and for those the opacity of the ink layer can only correspond precisely to the opacity of one of the individual inks: 8%, 18%, 30% or 44%. Some areas are printed by all four plates, producing a maximum combined opacity approximating 100% as four images overlap in the darkest areas. For shape elements composed of two or three plate images the potential opacity combinations become more varied, resulting in a total diversity of ink opacities that ranges from 8% up to 92%, with other values distributed in between (see Figure 8).

Although the hex elements in Figure 6 are of a similar exterior shape, their interiors contain different patterns of lines or cutouts that provide multiple density levels in the interior art of each hex. If artwork were designed with greater complexity than the simple hexes and semicircles in this example, it would even be possible to incorporate all fifteen opacity levels, and unique artwork, into the interior of every isolated shape in Figure 6. Contrast this with the basic monochromatic design strategy described in Part 2 of this series, which can only produce a number of density levels equal to the number of plates, if all fountains contain identical inks. A greater number of density levels allows for more design flexibility and more artwork complexity, and potentially a higher resistance to reverse engineering. The key question is whether it is possible to work backwards from Figure 6 to isolate its component images, Figures 2 through 5. Although Figure 6 could be originated and printed by a genuine document issuer, the design makes it hard for counterfeiters to deconstruct Figure 6 into its component parts, which prevents replication of individual plate artwork and the production of high-quality counterfeits.

Concealing the ink opacities
Varying ink opacity as described above can increase the number of discrete densities available, but does come with some potential drawbacks. For example, a few of the shapes in Figure 6 are printed from only one plate, which shows each individual ink in isolation and could help a dedicated counterfeiter to understand the artwork where multiple plate images do overlap. Allowing counterfeiters this insight is contrary to the stated purpose of the monochromatic strategy, which is to systematically deny access to information about the press setup and thereby prevent reverse engineering of the artwork.

There are three potential solutions to this challenge, all of which work to conceal not only the number of plates and the plate artwork, but also the opacities of the inks. One option is to minimise the amount of artwork printed only by one plate and maximise the amount of artwork composed of two or more over­lapping plates. However, this approach cannot fully eliminate small areas of single plate art inside the individual shapes (as shown in Figure 1, for example) without also causing undesirable design redundancies and reducing the availability of lighter opacity levels, so it is of limited utility. The second possibility is to use split fountains to transition between two inks of similar colour but of different opacities, instead of for the conventional purpose of transitioning between two different colours. This would allow multiple ink opacities to be deployed in each fountain, which prevents counterfeiters from associating a fountain to a single ink formula based only on examination of a small area of artwork printed by one plate. Split fountain ink opacity transition is a topic of considerable poten­tial that deserves its own analysis in a future paper of this series, so it will not be discussed further here. The third option, reducing visible opacity differences between the four inks in the set, relates directly to the subject of this paper and is addressed in more detail below.

Figure 8: Fifteen possible combinations created by overlap of inks with opacities of 8%, 18%, 30% and 44%. This set offers a wider range of combinations and more design flexibility. Compare to Figure 9.
Figure 9: Fifteen possible combinations created by overlap of inks with opacities of 18%, 22%, 27% and 33%. This set offers a more restricted range of combined densities, but the individual inks are harder to discriminate visually. Compare to Figure 8.

In the previous example, an association can be drawn between ink opacity and plate artwork because each fountain contains its own unique ink opacity. In theory, anything that helps a counterfeiter differentiate between the artwork on different plates could be exploited in a reverse engineering process. Therefore, the benefits of using inks of different opacities appear to conflict with the more fundamental goal to eliminate cues that could help a counterfeiter isolate the artwork on various printing plates. These competing priorities can be reconciled by choosing four inks with less disparate opacities that appear more visually similar, but can still be overlapped to generate a larger number of opacity combinations.The four individual inks shown in Figure 8 (8%, 18%, 30% and 44%) are easy to distinguish visually because they differ substantially in opacity. In contrast, Figure 9 shows a different set of four inks with subtly different opacities (18%, 22%, 27% and 33%) which are visually similar and more difficult to distinguish. For example, the 18% and 22% inks appear similar to one another, as do the 22% and 27% inks, and the 27% and 33% inks. The possible mid-level densities available in Figure 9 encompass a smaller range than those in Figure 8 (from 18% to 82%, instead of from 8% to 92%), so the use of more similar inks comes at the cost of some design flexibility, since the scope of middle values is reduced. Figures 10 through 14 illustrate how the more similar ink opacities shown in Figure 9 translate into a composite image that still contains fifteen different maximum densities, though they are constrained to a more compressed range.

Figure 10: Security design plate artwork printed with ink of 18% opacity.
Figure 11: Security design plate artwork printed with ink of 22% opacity.
Figure 12: Security design plate artwork printed with ink of 27% opacity.
Figure 13: Security design plate artwork printed with ink of 33% opacity.
Figure 14: Composite artwork from overprinting of the four plates shown in Figures 10 through 13, with a more narrow range of ink opacities than shown in Figure 6. Compared for Figure 6, some of the lightest and darkest middle values have been lost, but the individual plates may be harder to differentiate.

Compare Figure 6 to Figure 14. Both would be very complicated to reverse engineer, but each of these competing approaches has advantages and limitations. In Figure 6 the individual inks are easier to discriminate, which provides more design flexibility and a wider range of combination opacity values when artwork is overlapped. In Figure 14 the individual inks are harder to discriminate, which restricts the availability of some lighter and darker middle tones, but helps to further conceal the artwork on the individual plates. A hybrid approach, such as using a particular opacity more than once, is also an option (for example, opacities of 20%, 20%, 27% and 33%).

Conclusion
A limitation of the monochrome design strategy described in Part 2 of this series is its inability to produce a large number of discrete density values. Here, Part 3 has described how varying the opacity of inks within the monochrome design strategy enables a larger range of densities to be created using a fixed number of printing plates. However, associating specific ink opacities with specific printing plates could risk providing information to a counterfeiter about the press setup. To capture the benefits afforded by multiple ink opacities while reducing reverse engineering risk, a relatively narrow range of ink opacity values and liberal overprinting of plate artwork help mask the opacities of individual inks. Furthermore, the link between a single plate and a single discrete ink opacity can be broken by using split fountains to place multiple ink opacities in each fountain. This will be considered in future work.

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.

References

1 Zlotnick, J., Brough, J. and Eberhardt, T. (2016). Interrupting traditional counterfeiting workflow, Part 1: Colour and split fountains, Keesing Journal of Documents & Identity, Vol. 49, pp. 14-19.

2 Zlotnick, J., Brough, J. and Eberhardt, T. (2016). Interrupting traditional counterfeiting workflow, Part 2: Monochrome colour gamut, Keesing Journal of Documents & Identity,
Vol. 51, pp. 22-27.