Continuing our trilogy on optical machine authentication (previous parts please see Keesing issues 57 [1] and 58 [2]), this article focusses on future perspectives in this area. While the border guards of the German federal police have been the main users of document readers for more than a decade in Germany, other police and civil administration authorities are now initiating activities aiming to roll out document readers in larger scales in the federal states. Despite this attention, we anticipate that “conventional” inspection with the human eye will remain an important backbone in document examination for the next decades to come. We therefore take a look at security feature concepts facilitating both human and machine inspection processes. In fact, one of the aims of this article is to raise awareness of both document designers and inspection system manufacturers alike and to encourage the involved parties to invest in innovative security features and check routines considering this ambiguity of document inspection. However, before going into practical examples of such features, the major aims of the “EU Action Plan” as of late 2016 and the current status of corresponding activities regarding the aspects of document production and inspection within the action plan are discussed.
EU Action Plan 2016 and current situation in early 2019
The EU Commission adopted the so-called ‘Action Plan to strengthen the European response to travel document fraud’[3] in December 2016 with the aim to improve overall security in the EU by focusing on the entire chain of identity: registration of identity, issuance of documents, document production and last but not least, document inspection. The latter two aspects are fully in line with the topic of this article:
- Document production refers to the designing and manufacturing of secure, standardised and globally interoperable documents, of course mentioning that ICAO governs the standards for globally interoperable travel documents by setting out the specifications for (electronic) machine-readable travel documents (MRTDs and eMRTDs).
- Document inspection refers to efficient and secure reading and verification of travel documents. It also covers processes enabling a reliable linkage of documents and their holders to available and relevant data in the course of inspection.
Recent activities already address the aspect of document production, in particular the aspects of increased security and standardisation:
- Visa: a standardised and uniform format for visa with increased security (Regulation (EU) 2017/1370 of the European Parliament and of the Council of 4 July 2017 amending Council Regulation (EC) No 1683/95)[4]. For this new generation visa, a “Digital seal” in the form of a 2D barcode including a digital signature allowing the authentication of the personalisation content was planned. Unfortunately, however, the Council dismissed this feature: not all countries would have been able to introduce it right at the beginning of the introduction of this document.
- Residence permits: for third country nationals (Regulation (EU) 2017/1954 of the European Parliament and of the Council of 25 October 2017 amending Council Regulation (EC) 1030/2002)[5].
- ID-cards: on April, 17th 2018, the EU Commission adopted a legislative initiative to strengthen the security of identity cards of Union citizens and of residence documents issued to Union citizens and their non-EU family members exercising their right of free movement[6].
- Emergency travel documents (ETD): possibilities to modernise ETDs are explored and the EU Commission already adopted a legislative proposal introducing new rules with enhanced security features[7].
The European Border and Coast Guard Agency (FRONTEX) started several initiatives with the development of a methodology for testing and assessing the performance of Document Inspection Systems.
From 1998 the EU is developing the system “Expert FADO” (False and Authentic Documents Online) and its derivatives iFADO (internetFADO) and PRADO (Public Register of Authentic Documents Online). Since 2004, all member states can share descriptions and imagery of authentic and counterfeit documents via this platform. Currently, approximately 5400 data sets of document descriptions are available in FADO.
The newly started initiative called ProFID (Profiling of False ID Documents) deals with an IT-based information network, which collects trans-European information on counterfeit and falsified documents facilitating a comprehensive link analysis. Until the end of 2018 ProFID was already supported by twelve member states to perform testing under real conditions. EUROPOL is currently also considering to provide support.
Human vs. machine authentication – a contradiction?
In the previous section, the EU action plan has been discussed which found aspects of document production and document inspection to be of major importance for the overall security within the EU. Obviously, the main link between these two aspects, i.e. the production and the inspection of security documents, is the design and implementation of security features incorporated in such a document, which in turn facilitates its reliable and effective inspection e.g. at the EU borders. In this chapter, we discuss recent and earlier developments of security features carefully designed not only for inspection by the human eye, but also with the requirements and potentials of optical machine authentication processes in mind.Figure 1: Example of the primary facial image in a passport of Bulgaria (2009 series) –
left to right: plain datapage, datapage with lenticular lens on top, digitally calculated version using image processing to reconstruct the hidden image.
Figure 2: Example of the primary facial image in a passport of Portugal (2017 series) –
left to right: plain datapage, datapage with lenticular lens on top, digitally calculated version using image processing to reconstruct the hidden image.
Encodings in and within Photographs
One approach to implement machine-readable security features which has been around for about two decades is the encoding of hidden information using well-defined distortions in the rastering of printed elements. Such a feature could, for example, encode personal data (name, date of birth, document number, etc.) into the primary portrait. Thus, the printed photo, the document holder and the document itself are directly linked with each other through the encoded information. While the encoded data is invisible to the naked human eye, authenticity can be easily checked with a special decoding lens (human inspection) and/or digital image processing (machine authentication) as illustrated in Figure 1 and Figure 2.
Beside this readable information, digital data that is exclusively machine-readable could also be stored using the same approach: personal data in the form of a suitable 2D barcode could be hidden inside the photo with the corresponding decoding algorithm implemented in the software of the document reading device. This digital form of encoding would provide improved security over the previously described encoding as issuing states could even use their own encryption algorithms, if desired.
Furthermore, both features can even be embedded in the same facial image given that the orientation of the coding of both features is orthogonal to one another.
Figure 3: Examples of the secondary facial image in passports of A) SAR Hong-Kong (2007 series) and B) Finland (2017 series). Left the complete image and right the magnified view.
Beside these rather complex features requiring the use of a lenticular lens to uncover the hidden text in the photo, there are also other hidden features: The holder’s portrait, e.g. in a monochrome secondary facial image, can be comprised of horizontal microtext lines with personalisation content, again linking the portrait directly to the holder and the document itself. For authentication, only a magnifying glass is required in order to reveal the textual content of the microtext and compare it with the other personal data printed on the document (see Figure 3).
Recently, the next generation of this feature has been developed, aimed in particular at machine-assisted authentication: The secondary portrait is created along unique wavy structures, which are computed with a special algorithm incorporating the personal data of the document holder. Slight differences in letter size and line width create the necessary halftones for the monochromatic facial image. The manufacturer supplies the software to create this feature as well as a modular component which – once incorporated by the passport reader software – allows for the verification of the feature. This new feature effectively combines the possibility of simple visual inspection with machine inspection properties.
Figure 4: Example of a concealed secondary facial image in passports of Hungary (2006 series) and Portugal (2017 series) – left to right: images in visible light, infrared and under UV exposure.
Optically variable features with concealed individualisation
Another type of feature designed for both visual and machine-assisted inspection is by the use of translucent liquid crystal inks with optically variable properties: Located at the upper left corner of the personal data page of the 2017 Portuguese passport, a pearlescent optical-variable feature is created with liquid crystal inks on top of a dark background. In addition to the visible colour shift this feature has inherent polarisation properties that can be detected by standard polarisation filters or dedicated machine-assisted validators. Beneath the optical feature is a laser-engraved portrait that can be observed through an infrared viewer Figure 4.
Hence, one single feature provides multiple protection by addressing both visual inspection as well as machine detection.
Figure 5: Machine verifyable features “MPM” (inside the Identigram of German ID-Cards and Passports), “DAC” (inside the security thread of a German passport, 2017 series and inside the DOVID of the new 2018 EU-Visa label) and schematic verification geometry.
DOVIDs
The idea of machine verification of diffractive optically-variable image devices (DOVIDs) and holographic features dates back to the early years of the 21st century. Even though various types of complex and appealing DOVIDs are widely used in security documents, they remain difficult to reconstruct in optical machine authentication due to missing standards for document design and detection geometries. Despite this situation, the German ID-cards and passports developed in 2001 contain a feature called “MPM” (“Maschinenprüfmerkmal”), which would translate to something like “machine-verifiable feature”. The MPM can be recognised on German ID documents as the red circular-shaped dot in the Identigram layer just beneath the portrait. This type of feature can be authenticated by pointing a laser/led ray on its surface and verifying the resulting diffraction pattern inside the half-space above the document as depicted in Figure 5. The kinematic stripe on the back of the card utilises a similar structure called DAC (“diffractive area code”), which can also be found in the brand new series of EU-Visa labels and all previous versions since 2002.
When the MPM was introduced, passport readers built for machine authentication at German borders were equipped with a laser diode and a particular detection principle enabling the device to verify this feature. However, the device required moving parts inside the housing in order to illuminate the correct area with the laser diode, leading to extended inspection duration and reliability issues. Hence, this functionality was – unfortunately – removed for the subsequent models of passport readers of this manufacturer. The concept of this security feature, however, could still provide an increased level of security even at first line inspection, if a fast and innovative way of verification in document readers would be established.
Figure 6: The DOVID in the brand new EU-Visa label (Germany, 2018 series) – from left to right: images in visible light, infrared and under UV exposure.
In addition to the traditional DAC, the manufacturer of the DOVID (Diffractive Optical Variable Image Device) in the new EU visa stickers – issued from 2018 – went along another path to add security to the document using a novel technology suitable for machine inspection. With a proprietary production method, the feature in the DOVID is designed in such a way that metallised, optically active parts are in perfect register with printed parts. This unique feature can be verified by applying pattern matching algorithms in the visible, the ultraviolet and the infrared spectrally-selective images (Figure 6). The manufacturer provides corresponding software libraries to vendors of document readers for implementation which enable the corresponding check routines. Since the manufacturer itself defines the check routines, proprietary knowledge of the production processes including tolerances during manufacturing can be utilised which would not be otherwise accessible to the vendors of the document readers.
Philosophy
The impact and importance of optical machine authentication clearly increased during the last couple of years leading to a broader perception of the need for security features to be designed for both human as well as machine inspection. Some companies even went as far as to install a “dogma” that every new security feature being designed must have a balance between elements for visual inspection as well as machine inspection. We support this philosophy and encourage other parties to keep these ideas in mind when looking for innovative novel security features.
Such innovative ideas could, for example, also make use of the fact, that in document readers, the IR cut-off filter does not exist, leading to sensitivity for IR luminescent properties, thus, yielding a potential for novel authentication approaches even with existing hardware devices using tailored document designs with adapted verification algorithms.
Additionally, the ubiquitous presence of mobile ‘smart’ devices might open up new possibilities to enhance document security by utilizing the optical capabilities of such devices. Although IR and UV will be absent for most smart phones, varying angles and video analysis might enable new perspectives, which could compensate for this drawback. One could think of various features exploiting the optical possibilities of smartphones, e.g.:
- Detecting different spectral behaviour using the display as a light source with simultaneous acquisition of images of the front camera of the device, thus, mimicking “hyper-spectral” imaging techniques.
- Inspecting time-dependent properties such as phosphorescence using the high-power flash as a light source and the rear camera as a high-speed capturing device.
- Capturing a set of two images with a smartphone in two distinct lighting conditions and acquisition geometries: The first image is acquired utilizing the geometrically close vicinity of flash and rear camera, resembling a quasi-coaxial setup. The second image is captured in ubiquitous oblique daylight. The difference between the two images can then be analysed using image processing algorithms.
- The afore-mentioned typical property of smartphones with flash and rear camera being in close vicinity (quasi-coaxial lighting conditions) can also be utilised to verify retro-reflective security features as illustrated in Figure 7 for the Australian passport.
Figure 7: Images of an Australian passport (2008 series) acquired with a smart phone with (left) and without (right) flash. Due to the quasi-coaxial conditions, the image acquired with flash clearly shows the retro-reflective feature of the passport.
Summary
In this trilogy on optical machine authentication, a broad range of aspects surrounding this topic has been discussed:
In the first part[1], the rather general considerations summarised in the Best Practice Guidelines on Optical Machine Authentication part I[8] published by ICAO as well as the more specific ways for implementing information-rich logging of machine authentication processes and results (BSI technical guideline
TR-03135[9]) as a basis for insightful analyses and feedback cycles have been presented.
In part two,[2] we shared our experiences and difficulties within the project AROMA which aimed at evaluating machine authentication systems in a systematic and scientific manner. Even though this type of evaluation method clearly had its limitations, the valuable insights gained during the project will serve as a solid basis for future work on this topic. After having conducted the project IDEAL some years ago, AROMA proved (again) that – until today – hard- and software components of document inspection systems are not easily interchangeable or even compatible.
In this final paper, part 3, the concurrent EU activities in the field of document security and inspection have been discussed. Furthermore, practical examples of security features designed with machine authentication applications in mind have been provided. Even though human inspection remains important, we clearly see a trend towards such security features designed for both human and machine authentication and hope to see more innovative security features considering this ambiguity of document examination in the near future.
Tune out…
Acknowledgements
BKA would like to thank the governments and institutions of Australia, Bulgaria, Finland, Hungary, Portugal, Hong Kong and the EU for developing the security documents presented here whose careful design of security elements facilitates both human and machine-assisted document inspection.
We are also grateful to Klaus Emig for his assistance in acquiring the imagery for this paper.
References
- Weigand, C. and Schneider, U.,(2018). Optical Machine Authentication of Security Documents Part I: Recent International Impact, Keesing Journal of Documents & Identity, Issue 57, pp. 24-31.
- Weigand, C. and Schneider, U., (2019). Optical Machine Authentication of Security Documents Part II: Welcome to reality, Keesing Journal of Documents & Identity, Issue 58, pp. 7-13.
- Schneider, U. and Seidel, U., (2014). “Current Aspects in Machine Authentication of Security Documents – Part II: Unused potential and the need for improvement?”, Keesing Journal of Documents & Identity, Issue 43, pp. 3-12.
- Regulation (EU) 2017/1370 of the European Parliament and of the Council of 4 July 2017 amending Council Regulation (EC) No 1683/95 laying down a uniform format for visas, 04.07.2017. [online] Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32017R1370 [Accessed on 25.02.2019].
- Regulation (EU) 2017/1954 of the European Parliament and of the Council of 25 October 2017 amending Council Regulation (EC) No 1030/2002 laying down a uniform format for residence permits for third-country nationals, 25.10.2017. [online] Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32017R1954 [Accessed on 25.02.2019].
- Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL on strengthening the security of identity cards of Union citizens and of residence documents issued to Union citizens and their family members exercising their right of free movement, 17.04.2018. [online] Available at: URL: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri= CELEX:52018PC0212 [Accessed on 25.02.2019].
- Proposal for a COUNCIL DIRECTIVE establishing an EU Emergency Travel Document and repealing Decision 96/409/ CFSP, 31.05.2018. [online] Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52018PC0358 [Accessed on 25.02.2019].
- ICAO Best Practice Guidelines for Optical Machine Authentication Part 1: Recommendations. [online] Available at: https://www.icao.int/Security/FAL/TRIP/Documents/Best Practice Guidelines for Optical Machine Authentication V1.2.pdf. [Accessed on 25.02.2019]
- BSI Machine Authentication for Public Sector Applications, TR-03135, Version 2.3 2018. [online] Available at: https://www.bsi.bund.de/tr03135/. [Accessed on 25.02.2019]
Further reading
- Schneider, U. and Seidel, U., (2013). Current Aspects in Machine Authentication of Security Documents – Part I: Do we need optical document security?, Keesing Journal of Documents & Identity, Issue 41, pp. 3-10.
- Schneider, U. and Seidel, U., (2014). “Current Aspects in Machine Authentication of Security Documents – Part II: Unused potential and the need for improvement?”, Keesing Journal of Documents & Identity, Issue 43, pp. 3-12.
Christian Weigand received his PhD in electronics and telecommunication from the University of Trondheim in 2012. In 2016, he joined the Forensic Science Institute of the German Bundeskriminalamt, where he focusses on the forensic analysis of barcodes in identity documents, machine authentication and platforms for information exchange.
