Fight fakes with surface pattern recognition
Article By : Vivek Nanda
Fuji Xerox' Yoctrace technology uses image processing to recognise inherent random surface patterns on industrial products.
Remember the recent case of an Indian middleman colluding with a Chinese manufacturer to supply fake roller bearings for our guns? Or the fake new notes that immediately followed our famous demonitisation?
While EE Times Asia recently took a look at the scale of the problem worldwide in Counterfeits plague electronics industry, Fuji Xerox Co. Ltd announced a technology called Yoctrace that could help identify counterfeits in some applicaitons earlier and easier than current processes allow.
Yoctrace is an object recognition technology that recognises inherent random patterns inadvertently generated on the surface of objects, such as industrial products. The company has started licencing the technology.
The technology has been in development for sometime with the company first introducing a paper recognition technology back in 2002. Yoctrace builds on that concept and takes it to industrial materials other than paper enabling authenticity assessment and unique identification using an original algorithm.
The technology can be applied not only to the prevention of counterfeiting but also to the enhancement of security for saving virtual currencies on printed paper by combining with blockchain technology.
How it works
A scanner or something as simple as a smartphone camera is used to take the image of the surface patterns of a target product and the image is registered in the server. Then the image of the same part of the object that needs to be verified is captured and the two images are compared.
The accuracy of verification is high since the whole image containing random patterns is used for verification. This contrasts with conventional fingerprint authentication where only extracted characteristics of the fingerprints are used for verification.
Thus Yoctrace is fit to prevent the counterfeiting of gift certificates and ID cards for which a high level of security is required.
Figure 1: How it works. Fuji Xerox claims the verification of surface images is not be affected by color shading or unevenness of object surfaces when taking pictures.
No tags
Since the target of verification is an image of the object itself there is no need for tagging individual objects using serial numbers or barcodes, which allows for the identification of small objects where tagging is difficult, such as pills.
With the aim to enable application in various industries, Fuji Xerox is undertaking further research to identify the uniqueness of an object—not only to prevent counterfeiting, but also as an IoT technology—linking individual history of various objects with their manufacturing history to aid quality and logistics management.