App Claims It Can Spot Fake Products With Just an iPhone Photo
FeaturePrint is a new artificial intelligence (AI) powered app that developer Alitheon claims is able to see a product and determine if it is faked through just a single iPhone photo.
Alitheon says that faked products are a major problem and the issue is even expanding into the art world and to auction houses. While there are experts who specialize in recognizing fakes, they can still be fooled.
“Auction houses are under pressure to ensure that the art they sell is authentic. A single incident of selling a fake could damage their reputation irreparably,” Alitheon’s public relations says. “In fact, according to the United Nations Office on Drugs and Crime, the underground art market, which includes thefts, fakes, illegal imports, and organized looting, may bring in as much as $6 billion annually.”
The company says that it has developed a simple solution that makes identifying fakes extremely simple. Using only a single photo taken with an iPhone, FeaturePrint — an Optical AI — is able to “see” and convert the surface details of physical items into what Alitheon describes as a “unique mathematical identity.”
The company claims that because no two objects are the same even when they come out of the same production line, the app can “easily identify genuine products through the unique digital fingerprint inherent in each item.”
While the app uses a photo to analyze a product, the company makes it clear that FeaturePrint isn’t using an image to identify an object.
“A FeaturePrint is a set of numbers characterizing the unique features of an object. Features common to similar objects are rejected, and only the information that makes this object unique is preserved,” the company explains.
What’s more, the app characterizes the speific object from which it was extracted, not the kind of object.
“It may, for example, represent a particular Intel Core i9 processor, not the class of all Intel Core i9’s. All class-based information is intentionally removed,” Alitheon says.
It says that this technology is about machine vision and optical AI algorithms, not stickers, bar codes, QR codes, RFID, or etchings.
“Most identification systems today use proxies such as barcodes, QR codes, or RFID tags. These can be obscured, detach from the object, or themselves be counterfeited. By making the object its own identifier, Alitheon does not need them,” the company claims.
“Most of the world’s goods are either untraceable or only known through an identification proxy added to objects. Barcode labels, etchings, seals, exotic materials, are variously used to declare the provenance of individual items. They can also be lost, damaged, faked, changed, and manipulated,” Alitheon continues.
“Should an item’s proxy be lost — or never exist — the product is suddenly anonymous. Its performance, safety, or economic value is unprovable. In global B2B and B2C commerce, such failures cost hundreds of billions of dollars annually due to scrap, recalls, damage, and theft. Customer trust deteriorates with every loss.”
The company says that this technology is a “game-changer” for the art and auction world as it helps reduce the risk of fraud and builds trust between buyers and sellers.
It should be clear that FeaturePrint can’t actually determine if a product is a counterfiet, but rather can only confirm authenticated items that have been registered to the system.
FeaturePrint isn’t available through smartphone app stores and Alitheon provides it as a business-to-business solution with customers directly. It is, however, already in use by FAST Sneaks. At the time of publication, the company doesn’t list any pricing or information on how to use the software on its website.
Update 5/26: Clarity added.