Counterfeit Detection Scheme using Paper Surfaces and Mobile Cameras

Abstract

ISR-affiliated Professor Min Wu (ECE/UMIACS) and alumnus Chau-Wai Wong (EE Ph.D. 2017), are the inventors on U.S. Patent US 10,726,256, granted July 28, 2020.

“Counterfeit Detection Scheme Using Paper Surfaces and Mobile Cameras” is based on Wong’s dissertation research. Wong is an assistant professor in the Electrical and Computer Engineering Department at North Carolina State University. At Maryland, Dr. Wong was jointly advised by Min Wu and Gang Qu (ECE/ISR).

Merchandise packaging and valuable documents such as tickets and IDs are common targets for counterfeiters. Traditionally, counterfeiting has been thwarted by adding high-cost surface structures such as holograms, ultraviolet ink, and random colored fibers. These techniques tend to be expensive and have a weak ground truth. In addition, authenticity detection relies on a manual decision made by the recipient of the item.

As an alternative, the optical features of low-cost surface structures can be exploited for counterfeit detection. Surface structures can be added with fiber, small plastic dots, air bubbles, powders and glitters foreign to the surface.

But even the randomness of a paper surface can make the structure physically unclonable or very difficult to clone The uniqueness of the inherent 3D structure of paper can be exploited for authentication purposes.

Wu and Wong’s patent is a method that uses the optical effect of the microscopic roughness of paper’s surface for authentication.

Using mobile cameras and illumination—such as a cell phone’s “flash” feature and camera—an item’s surface is illuminated and captured in multiple images. The camera-surface geometry is predefined by estimating the location of the camera lens and calculating a direction of incident light for every pixel of a surface area. Then, the item is authenticated based on comparing the captured images to a reference appearance that uses the same predefined camera-surface geometry. The authentication is based on a microscopic normal vector field of the surface based on the direction of incident light. With high enough camera resolution, collecting normal vectors of the physical paper surface can serve as a fingerprint that authenticates the item.

U.S. Patent and Trademark Office Description

PTO

Inventor(s)

Min Wu, Chau-Wai Wong

Date Issued

07/28/2020

Patent No.

10,726,256