News
No chance for forgers
Forging official documents is a billion-dollar business. Now, researchers at the ETH Zurich Centre for Cyber Trust have developed a system that uses digital technologies to authenticate paper documents—and the city of Zurich has already begun using the new method.
Birth certificates, family records, business contracts, university diplomas: although digitisation has advanced quickly, many important records are still printed, sent and stored on paper. With all the advantages and disadvantages that go with it. One of the disadvantages is that paper documents can be forged. In 2015 alone, a company in Pakistan is believed to have sold more than 200,000 fake academic degrees, titles and transcripts for upwards of fifty million Swiss francs.
Stamps, seals and signatures are common ways of validating the authenticity of a paper document. “However, for most recipients, it’s impossible to know whether a stamp or signature is real,” says David Basin, head of the Institute of Information Security at ETH Zurich and co-leader of the Centre for Cyber Trust, which the Werner Siemens Foundation finances.
The sure way to find a counterfeit is by comparing a document to the original version. But this isn't always possible, and even when it is, the procedure is time-consuming, as forgeries are often very sophisticated and thus difficult to detect. To make the process easier, a research team led by David Basin has now developed a method to easily authenticate paper documents using digital means. The researchers presented their study (1) last December at a conference in Austin, Texas.
Comparison using a smartphone
The core part of the new method is a smartphone app that can make a video of a document and then compare the video sequences with an encrypted original or reference document that is stored in a safe location. If it discovers any relevant differences, the app alerts users, who consequently decide whether the document is trustworthy or whether it’s a fake.
Basin explains that the concept is based on two key computer science procedures. The first is cryptography: encryption and digital access control protect the original document from unauthorised access to the storage site; in addition, digital signatures ensure that users download the right reference document from the right server.
In the second key procedure, image processing algorithms are used to compare the digital reference document to the paper document that a user wishes to authenticate. Although advanced automated text recognition programs have already been developed, as Basin explains, these solutions function on the basis of recognising text characters, but not layout elements, graphics and non-standardised symbols. “In contrast, our solution is based on comparing the entire image.”
Problem points: creases and stains
An iterative principle is at the heart of the new algorithm: if the app finds a potential difference between the two documents, it zooms in on the specific place. “It’s a little like a person using a magnifying glass to take a closer look at a suspicious section,” Basin says. That’s why it’s an advantage that the app makes a video of the document and not just a single image: because a person’s hand moves while making the video, the resulting images are all taken from slightly different positions. “The app delivers various pictures, which significantly improves accuracy,” as Basin explains.
The researchers conducted in-depth tests on their method using nearly 1000 forged documents and several smartphone models. The algorithm identified every single counterfeit. Basin then explains that the greatest challenge in comparing documents lies not in recognising fake documents but in dealing with unwanted changes. For example, documents are often folded when sent by mail, and sometimes coffee is accidentally spilled on them or the app is used in poor lighting conditions.
In such cases, the app indicates the differences, and users can adjust a slider to compare the two versions—and thus ascertain themselves whether the document is indeed a fake, or whether the software raised a false alarm. “Our tests have shown that this is a very intuitive way for users to receive and process this kind of information,” Basin says.
City of Zurich implementing the method
The innovative detection software has met with great interest. Basin says that the city of Zurich introduced the solution at the end of March for authenticating various documents, with the initial phase focusing on excerpts from the debt collection agency. Persons asking for an excerpt receive a paper copy by regular mail, as is the current process, but they also receive a QR code to access the original document; they can then download a free app to their smartphone to compare the copy with the original.
The new software has good potential for testing many other types of documents, Basin says. At the municipal level, this would include vital records such as marriage, divorce, birth and death certificates. But professional diplomas, bank documents and business correspondence are also at risk of being forged. “Fraud is a billion-dollar business and a large part of it comes down to falsified communications,” David Basin explains. “If we succeed in getting just a small part of this under control, we’ve made a major contribution to creating a safer society.”