Thankfully, artificial intelligence is here to help. A new machine-learning system developed by researchers from New York University is able to distinguish between genuine and counterfeit products by analyzing microscopic characteristics that are invisible to the human eye.
“We built Entrupy as a scalable and versatile platform in response to the rapidly growing counterfeiting issue and need for trust when it comes product transactions,” Vidyuth Srinivasan, co-founder of Entrupy, the company that has commercialized the technology, told Digital Trends.
The non-intrusive Entrupy system uses a dataset of 3 million microscopic images, including goods and materials like fabrics, leather, electronics, toys, and shoes.
“Entrupy’s technology is a mix of machine learning and microscopy,” Ashlesh Sharma, Entrupy’s fellow co-founder, told Digital Trends. “We train our machine-learning algorithms to pick up data points from millions of microscopic images looking for qualities like texture, contrast, topology, geometric shapes, thread counts, minor manufacturing artifacts such as scratches in the hardware stamps, wear, and many more details that you wouldn’t be able to easily see. These details are in fed into our custom machine learning pipeline, allowing us to determine a product’s authenticity. ”
Sharma said the system is currently 98.5 percent accurate but that, as a machine-learning system, it is improving with every use. “With machine-learning technology, our algorithms are always getting better, building a better database of what makes a product authentic, and even more importantly, what details mark a counterfeit,” she said.
There are, of course, other methods to distinguish genuine and counterfeit products, but they are generally invasive and may end up damaging the product. Entrupy boasts that all it needs is its scanner and image database, which do not interfere with the product itself.
As Entrupy’s price suggests, the device is not intended to be sold to the individual shopper — unless you happen to be a big spender. Rather, it is a way for retailer and wholesalers to make sure they are selling the real thing, offering a certificate of authenticity to customers. The system and scanner cost $99 for five scans per month, $399 for 30 scans, and $999 for 100.
