Automated examination and authenticity control of secure holograms.
Regula began to look at the automated examination of DOVIDS a number of years ago and produced a working prototype. This was then followed by further developments after an example of the need for such a system was shown at the 2014 ODS conference--during which a presentation on a fake permanent resident card that incorporated a counterfeit University of California laminate containing an eagle design and text was shown.
So Regula began to investigate the security features of different diffractive optically variable image devices (DOVIDs) and the tools and examination methods used.
Two DOVID examination methods typically used were:
* The 'image in a hole' approach, where the object DOVID image is observed and any changes compared to the known authentic DOVID image;
* A second approach that uses devices such as lasers or microscopes to examine and measure grating structure parameters of the DOVID.
Other methods that use--for example--mobile phones or a fixed number of light emitting diodes (LEDs) produce a general overview of the DOVID but can be limited in DOVID visualisation.
What affects DOVID visualisation?
Regula created an algorithm that generated the intelligence to be able to distinguish between real and fake DOVID images (image comparison). In principle, this may sound easy, but in reality is difficult as factors such as curvature of the substrate to which the DOVID is applied, hot stamping accuracy and the position of the DOVID during examination can all affect its visualisation.
Other factors that also have to be considered in any image comparison algorithm are creases or folds in applied security DOVIDs, as these also affect how they are visualised.
An example is shown for the Schengen visa hologram after being creased. 50% of the surface has a different optical response from a crease that is less than 3% of the whole surface.
The number of illumination points is also an important factor affecting the visualisation, as a much wider spectrum of reference points are present for 360 illumination positions compared to 32 illumination positions.
Regula's analysis of existing DOVID examination methods found a correlation between the complexity of the DOVIDs and the examination methods used. The microscopy or laser based examination methods provide high quality and accurate analysis, but the time of analysis depends on the area of the DOVID analysed. Conversely, both mobile phone and LED examination methods provide good (satisfactory) time of analysis regardless of the area analysed.
Regula established that using the LED examination method can provide a comparable analysis quality to laser system analysis. The verification time for the LED method is also independent of DOVID size. However this method depends on the complexity of the DOVID as a larger number of scanning angles are required during analysis.
Regula's objective for the development of an automated examination and authentication method for holograms was twofold:
1 To expand the use of LED's to complex DOVIDs;
2 Achieve examination time of less than 10 minutes.
In order to achieve these objectives, Regula took the approach of using the DOVID's inherent ability to diffract white light into its component colours with varying the zenith angle (angle of illumination). The variation in this angle causes a change in the wavelength and hence colour pixel that is detected by a camera.
Similarly, changing the period of the DOVID's gratings when using a fixed light source at a fixed angle results in a different observed colour for the illumination.
Using the above approach together with calculations allows a map to be created, enabling the parameters of the DOVIDS gratings such as the angle and position to be calculated for every point of the DOVID under examination.
This is a difficult task, is time consuming and involves taking a known DOVID (a reference) to create a reference matrix or raster point for every pixel on the DOVID that is then used as a reference system for the DOVID under examination.
The algorithms can then be applied without having, in theory, to scan the entire DOVID in question, as reference points can be used which are critical for the decision making process.
However, there still remains the issue of the 'homogeneity map' which takes into account every position and orientation of the DOVID in question when compared to the reference. Some of the examined DOVIDs may not be completely flat, or are skewed, and some of them may be rotated. All of these possibilities have to be taken into account and recognised by the artificial intelligence when the DOVID is being examined.
So far, provisional tests indicate that the system works once the reference matrix of the DOVIDs are calculated and known. No full 360[degrees] examinations are necessary for examination and authentication of the DOVIDs. This enables the use of mobile phones that incorporate the algorithms, which results in much faster examination times.
How would it work in practice?
The algorithm for the analysis would be bespoke per DOVID and retrieved from a database. For example, the inspector would perform a quick scan to identify the object (eg. a Kinegram from a banknote), and would then select the kinegram-specific algorithm, select the reference to be applied, apply the homogeneity maps ensuring all the inaccuracies in the positioning are included, and then apply the algorithms to obtain the result.
Alternatively, a quick scan could be performed to give a yes/no type analysis and, if needed, an entire 360[degrees] scan to provide a forensic analysis of the DOVID. The latter analysis would be secondary and justified by previous examinations to allocate the time if required.
Using the above approach has enabled Regula to develop a commercially-available desktop device for automatic recognition of DOVIDs. The device has a number of zooms (max and min zoom), a maximum resolution of 27,500 dots per inch, and can measure a number of parameters including, but not limited to, the period and orientation of the diffraction grating.
The device has a number of advantages including:
* Fast and efficient evaluation of the entire surface of DOVIDs;
* The capability to detect forgeries at the level of grating parameters;
* Use of white light and not microscopy tools.
For the future, Regula is continuing to further develop the software, reduce the time required for DOVID examination and introduce additional features.
Caption: An example of a mobile phone examining a DOVID on a Euro banknote.
Caption: Left image shows the same DOVIDs applied to different surfaces of curvature. Right image shows how the position of the DOVID affects visualisation.
Caption: Schengen visa DOVID visualisation viewed under identical viewing conditions; the top image has no fold or crease in the DOVID whilst the bottom image has a fold in the DOVID.
Caption: Identical DOVID illuminated with different numbers of LEDs. Top image is DOVID illuminated with 32 LEDs, while bottom image is DOVID illuminated with 360 LEDs.
Caption: Time versus quality of examination for different verification methods of DOVIDs.
Caption: Diagrams of the variation of zenith angle.
Caption: Proposed authenticity approach.
Caption: Regula automated DOVID desktop examination and authenticity device.
Please note: Illustration(s) are not available due to copyright restrictions.