Manipulating digital images isn’t anything new, in fact it goes way back in history. After smoking was deemed unhealthy, the cigar in this image was removed from Winston Churchill’s mouth:
And then there’s the more modern, certainly more obvious digital image manipulation we all love…
But, the issue of digital image manipulation in the identity verification space is certainly no laughing matter. Modifying part of an identity document can take just minutes using a photo editing software such as Photoshop, and more than often than not goes completely undetected by the naked eye, opening your business up to higher risk of fraud and loss.
Take this driving licence for example. You can barely even tell it’s been digitally altered to show a different date of birth:
But that’s where digital image tamper detection technology helps…
But how does it work? Let’s get technical…
A typical camera has several components as shown below:
These components leave different kinds of artefacts (fingerprints) that can later be used to detect image tampering.
Examples of artefacts include:
- Chromatic aberration
- Sensor noise
- JPEG compression
When an image is digitally altered using a photo editor, these artefacts become inconsistent, and the digitally manipulated area will exhibit different artefacts when compared to the other, genuine, parts of the image. Digital image tamper detection technology uses mathematical methods and machine based learning to automatically detect and localise these parts of an image that have been digitally altered.
When a JPEG image is altered, typically it’s decompressed, loaded into photo-editing software, manipulated, and then recompressed. Digital tamper detection technology detects the changes that JPEG compression makes to the image. You can see above that the modified areas have a different JPEG processing history and pattern in comparison to genuine areas of the image, showing evidence of digital tampering.
Detecting medium of document images
Smartphones have made storing and sharing physical documents easier than ever. However, with no access to the physical document itself, the question is whether you can recognise the medium of the document. It’s critical for remote identity verification that the medium of the document is understood, and whether it represents a picture of a genuine and government-issued identity card, passport or diving licence.
Digital tamper detection technology automatically detects the print quality and medium on which the document has been printed, with a single picture. Using AI, the technology performs analysis of fine details, differentiating pictures of identity cards created by printers used by governmental authorities on genuine materials, and those that are printed by fraudsters.
The image below shows a genuine document (top), and the same printed by a casual counterfeiter (bottom). Casual counterfeiters often rely on ink jet printing to mimic or simulate the genuine document’s artwork.
Carrying out visual checks of your customers’ ID documents just isn’t enough to protect your organisation against fraud. Digital tamper detection technology goes well beyond what the naked eye can see, to give you the evidence you need to make the right decisions about your customers. Distinguish your genuine customers from the fraudsters, and protect yourself against loss.