Designed, developed and implemented a novel technique for encoding data into and decoding
data from a coded pattern using an angular symbology in which data symbols are represented
by angular orientations of data modulation patterns relative to a reference modulation
The modulation patterns are selected to have components that are localized in a Fourier
transform domain and the technique may convey arbitrary digital data within 2D images.
This symbology's targeted regime of operation would typically involve D/A and A/D stages:
a coded image is printed on some physical medium, and subsequently optically re-digitized
prior to the retrieval of the embedded digital data. The technique was explicitly designed
to minimize any registration issue, which typically occurs for this type of technology.
It was also designed to be graphically versatile and sensitive to the appearance of the
resulting coded images. Additionally, it exhibits the unique ability to convey positioning
information throughout the image that carries it. Not only can the data payload be retrieved
from any section of the coded printout, but the position of the imager with respect to the
printout can also be estimated.
The modulation patterns are selected to have components that are localized in a Fourier transform domain and the technique may convey arbitrary digital data within 2D images. This symbology's targeted regime of operation would typically involve D/A and A/D stages: a coded image is printed on some physical medium, and subsequently optically re-digitized prior to the retrieval of the embedded digital data. The technique was explicitly designed to minimize any registration issue, which typically occurs for this type of technology. It was also designed to be graphically versatile and sensitive to the appearance of the resulting coded images. Additionally, it exhibits the unique ability to convey positioning information throughout the image that carries it. Not only can the data payload be retrieved from any section of the coded printout, but the position of the imager with respect to the printout can also be estimated.
Escher Labs' SpectraSeal technology is a high density, visually pleasing 2D symbology. It is an elegant alternative to symbologies such as 1D or 2D barcodes, allowing a computer to embed arbitrary digital data into a 2D representation. The symbology does not place any restrictions on the size, shape, or color of the image into which data is encoded.
Figure 1 - Three examples of SpectraSeals illustrating the wide range of possibilities offered by Escher Labs' novel 2D symbology. All three images contain the same data, which can be retrieved by the same reader.
Escher's SpectraSeal has been shown to work very well as a counterpart to Escher's FiberFingerprinting analysis. Indeed, the same 2D imager that is required for a fiber analysis stage can read the SpectraSeal efficiently. Also, its integration within any arbitrary graphic design is unmatched by any other symbology (see Figure 2 as an illustration). The SpectraSeal has been shown to be robust against severe physical damage of the medium it's printed upon, ranging from punctures to scratches and even scribbles.
Escher's SpectraSeal symbology was designed with a general mode of operation in mind. The symbology's robustness (through printing, degradations, and scanning) is an obvious concern. Registration and alignment are also recurring difficulties when it comes to printed symbologies and handheld readers. In addition, the printed symbology's visual aspect (shape, size etc.) can play an important role for applications where great efforts where put into graphic design.
These general concerns have guided Escher during the design of its SpectraSeal technology. The result is a symbology that is particularly well adapted to applications where a fixed amount of data is embedded in a variable amount of printable area in a robust and graphically versatile manner. This preferred mode of operation is reflected by the SpectraSeal's intrinsic properties
Non-locality of the data
Escher's SpectraSeal conveys data in a spatially distributed manner. This particular design choice was motivated by the decoder's registration issue, but more importantly by robustness concerns. As each bit of the payload is spread over the physical area of the symbology, the payload as a whole is able to cope with mechanical degradations such as scratches, punctures or stains.
Registration-free by design
From a visual capture of a printed area, estimating the amounts of translation, rotation and scaling in order to perform a precise alignment can be a rather computationally expensive and unreliable proposition. This is especially the case if one wishes to accommodate a large range of graphical aspects for the symbology.
For this reason, Escher's SpectraSeal was developed to be registration-free by design. The reading process is oblivious to any rotation or translation that might have occurred, making it particularly well adapted to handheld applications.
Escher's SpectraSeal can accommodate a wide range of visual constraints. It can be combined with an existing image or printed by itself. There is virtually no restriction over the shape of the resulting 2D representation. It can make use of color or grayscale but it can also accommodate a black and white channel, which could be imposed by the use of thermal or dot matrix printers.
Robustness is a measure of the embedded data's persistence through degradations of the paper medium the SpectraSeal is printed on (i.e. the channel).
In the context of more traditional channels (radio frequency for instance), a communication system's robustness is often measured in terms of the data's persistence through decreasing signal to noise ratios (SNR). The relevance of such measurement comes from the fact that the typical degradation of the channel can be reasonably modeled as an additive white noise. In that case, the ratio of the carrier and the additive noise energies is a justifiable quantitative measure of degradation.
Unfortunately, this methodology does not translate well to our case, where a printed paper medium is the communication channel. Indeed, among the large amount of possible types of degradations that may be applied to a piece of paper, very few could be reasonable modeled as an additive white noise. So while it may be possible to derive a similar SNR measurement of a printed symbology's robustness, the relevance of such measure would be greatly questionable.
Figure 2 - Example of some severe physical degradations applied to a printed SpectraSeal. The embedded data (106 bits total payload) is retrieved effortlessly from any of the image captures shown above.
Figure 2 illustrates a few types of physical degradations that are likely to occur in the real world (ink rubbed out, scribbles and punctures), none of which sharing much similarity to an additive white noise. Despite the severity of these examples, the 106-bits total payload of the shown SpectraSeal survives these degradations.
In the light of the highly configurable nature of Escher's SpectraSeal, it is not surprising that the data density of this symbology does not boil down to a single number of bits per square inch measurement. The data density of a SpectraSeal will always be a function of the tradeoffs that took place during the design of a particular application.
Because of the intrinsic nature of the SpectraSeal symbology, namely the non-locality of the data is carries, its mode of operation is better explained as the encoding of a fixed amount of data over a variable amount of area, rather than a fixed area per data symbol. This is a departure from symbologies such as traditional barcodes, which allocate a fixed area to a single data symbol. However, one can still derive a similar "bit per area" data density measure C (bit.in-2) for a symbology such as Escher’s SpectraSeal, by considering the minimum area Amin (in2) that is required to convey a fixed amount of data L (bits).
The most straight forward limiting factor for the symbology’s data density is the lower of the printing and scanning resolutions that are used for the application. It can also be finely tuned in accordance with specific tradeoffs between robustness and graphical considerations. In the context of various applications, Escher has used different configurations of its SpectraSeal symbology, spanning a wide range of effective data densities from 300 to over 8,000 bits per square inches.
Escher Labs' FiberFingerprinting technology is a means to produce secure documents from arbitrary (potentially low-value) stock, using commercially available variable printing technology. By applying digital technology to a physical medium, an Escher FiberFingerprint can authenticate a document with a level of security that surpasses that of traditional techniques such as microprinting, pantographs or holograms.
Figure 4 - Sample application of the SpectraSeal as a counter-part to Escher's Fiberfingerprint analysis
To verify the authenticity of a document, an imager checks for a match between the physical substrate (using Escher's FiberFingerprint analysis) and the printed data (potentially using Escher's SpectraSeal). If the match fails, the document is revealed to be a counterfeit. Thus, the symbols printed on a document can be copied indefinitely, but only the original will be valid. This technology allows a variety of value documents to be produced and generated electronically. These include money orders, checks, stamps, proofs of purchase, coupons, and legal documents. Unlike other technologies for identifying documents (e.g. radio-frequency identification tags), FiberFingerprint exploits the intrinsic nature of the medium itself. The cost of employing it is low, however, as commonplace printers and imagers can be used to produce and verify these symbols.
This application of the SpectraSeal demonstrated its reliability, its ability to accommodate low resolution, black and white impact printing (see Figure 4), and its ability to accommodate a large obstruction (the Fiberfingerprint area and registration mark).
For another project, the versatility of the SpectraSeal was further demonstrated with a token identification demonstration. Here, SpectraSeals conveying unique identifiers were engraved on plastic tokens.
Figure 5 - Token Identification based on an engraved SpectraSeal
The engraving was accomplished using a laser cutter set to appropriately low power. The identity of each token was then reliably recovered when placed on top of a 2D imager, once again demonstrating the SpectraSeal's reliability and ability to accommodate a wide range of "printing" technologies.
The premise of this application is to convey a unique identifier on a postage stamp for track and trace purposes. Given the nature of the medium, this should be done in such ways as to minimize any visual artifacts associated with the embedding of the identifier.
Figure 6 - Two samples of traceable stamps using subtle SpectraSeals to convey their identity
The unique identifier of a stamp can be pulled out reliably using a consumer-grade 2D imager. This application mainly illustrates the ability for a SpectraSeal to operate in a near imperceptible manner. This is particularly important in cases where graphic design plays an important role in the nature of the product.
The SpectraSeal is a high performance symbology. It persists through severe degradation of the physical medium that carries it: obstructions, scratches, overlay, etc. It offers a high data density in the sense that it can recover a significant payload from a small physical area.
The SpectraSeal is a versatile symbology. It supports a wide range of printing constraints in terms of resolution, color depth and quality. It also supports a wide range of imaging constraints in terms of field of view, color depth and quality. And finally it supports a wide range of graphic design constraints as it can take virtually any shape and size, and be combined in a near-imperceptible manner with an arbitrary image.
As a result, the SpectraSeal affords a wide range of applications, from overt to covert encoding of identification data, security data, or other meta data.