An Indian-origin research team at New York University has created an artificial intelligence (AI) system coupled with a smartphone app that can spot the difference between a genuine and a fake product.
Using over 3 million images of fabrics, pills, leather, toys, shoes and electronics, the AI algorithms are able to identify a fake product more than nine times out of ten.
“The classification accuracy is more than 98 per cent, and we show how our system works with a cellphone to verify the authenticity of everyday objects,” said Lakshminarayanan Subramanian, Professor at New York University.
The AI system will be showcased at the KDD Conference on Knowledge Discovery and Data Mining in Halifax, Nova Scotia, Canada, on August 14. As part of this initiative, a startup founded at New York University, called Entrupy Inc., will commercialize the product, which essentially works through a iOS app for both iPhone and iPad.
“The underlying principle of our system stems from the idea that microscopic characteristics in a genuine product or a class of products – corresponding to the same larger product line–exhibit inherent similarities that can be used to distinguish these products from their corresponding counterfeit versions.”
So, while there may be minor differences in quality within a particular genuine product line, there are similarities at the microscopic level that the algorithms pick up on to identify whether or not a product is genuine.
There are several methods to detect fake products, but most are intrusive and require invasive techniques that can often damage the product being analyzed. The Entrupy method using AI algorithms, on the other hand, is non-intrusive since it primarily uses imagery that is compared with a known dataset of millions of images.
Entrupy Inc. was founded by Ashlesh Sharma, Vidyuth Srinivasan, and Laxminarayanan Subramanian, who will be presenting the product at the Canada event tomorrow, August 14.