NEW YORK (Reuters) - In a designer-obsessed world, counterfeiters run amok.
New York start-up Entrupy has found a way to fight the fakes with a handheld microscope camera that lets users detect the difference between real Gucci, Chanel, Hermès and Louis Vuitton luxury handbags and frauds, with their smartphones.
"Nobody's ever done this before right? So we had to catalog the world in a microscopic way and first understand if there's actually a difference," said Vidyuth Srinivasan, co-founder and CEO of Entrupy.
Srinivasan and his team spent two painstaking years gathering data from all over the world. From buying and returning luxury goods at retail stores to imploring friends and second-hand shops to let them scan the items into their database.
"It works with a mix of microscopy and machine learning, or AI, as it's called today. What we do is we essentially collect a massive - I'm talking tens of millions of microscopic images - from physical items. So in a sense we're cataloging the world of physical objects in a microscopic way and then we're teaching computers to understand the differences in how authentic objects are made, versus how fakes are made," added Srinivasan.
Srinivasan said Entrupy has a 98.5 per cent accuracy rate.
"You know you can't say that, 'Hey this Rolex watch is about 80 per cent authentic', it's either real or it's not," he said.
In only a year since Entrupy's launch to the public, they've acquired 200 clients made up of businesses that include second-hand shops and online retail stores.
The device is leased for an initial fee of US$299 (S$408.41) with monthly plans starting from US$99 and can detect from 11 brands.
Srinivasan said his handheld microscope camera, which scans the item by taking photographs of multiple regions then magnifying the image 260 times, is meant for businesses for now, and not really meant for an individual consumer.
"Some of our customers see their business grow 50, 60 per cent. Some of our customers have quit their day jobs and now are selling luxury handbags online as a business. So, if we're empowering consumers and businesses, I think our job is halfway there," said Srinivasan.
Could counterfeiters potentially game the system by taking advantage of Entrupy's device?
"We collect millions of images across every single skew. So the depth of data that is available within the database that we use to classify and train the computers is way more than what counterfeiters would have access to, even if they were to use our device to keep trying to check and match. Ultimately it comes down to the fact that if you are not the manufacturer and if you're not manufacturing in the exact same way with the exact same materials, exact same workmanship, and even something as detailed as the depth of a hardware stamp, then you are going to be detectable," he said.
Counterfeit and pirated goods accounted for up to 2.5 per cent of world trade, or as much as US$461 billion, according to last year's figures from the OECD (Organisation for Economic Cooperation and Development).