A team of researchers led by New York University professor Lakshminarayanan Subramanian has developed a new method that uses machine learning algorithms to spot counterfeit consumer products.
As Subramanian explains, the underlying principle of the new method comes from the idea that genuine products have certain microscopic characteristics which correspond to the same larger product line. These characteristics can be used to distinguish real, genuine products from their corresponding counterfeit versions.
Counterfeit products are a global problem with pretty much every high-valued product being directly affected by this issue. In fact, some reports suggest that counterfeit trafficking represents 7% of the world trade today.
The new method provides a non-intrusive solution by deploying a dataset of three million images across different objects and materials, including fabrics, electronics, pills, shoes and even toys. The machine learning algorithms that are used provide a classification accuracy of more than 98%, which is the best by now.
The system works via cellphones – so soon, using your smartphone, you’ll be able to verify the authenticity of your favorite products and other physical objects.
New York University via Phys.org (https://phys.org/news/2017-08-machine-counterfeit-consumer-products.html)