Communications of the ACM
niversity Enhances Visual Identification Technologies For Retail Applications
This 3-D model of a store was created by a five-minute scan using imaging technology commonly found in cellphones.
Credit: University of Arkansas
The RFID Research Center in the Sam M. Walton College of Business and the Center for Advanced Spatial Technologies (CAST) in the J. William Fulbright College of Arts and Sciences at the University of Arkansas have partnered to conduct research that will integrate emerging visual identification technologies to use for retail applications.
The two centers will also work together to develop standards for widespread adoption of the technologies.
Visual identification technologies (VIT) use inexpensive 2-D and 3-D imaging technologies, commonly found in cell phones and video game devices, to identify objects by color, shape, and size without the need for traditional barcodes or product numbers. Products on store shelves or in warehouses can be recognized quickly and linked to product information databases. When linked to associated product information, the product can be added to inventory, verified for correct location, and removed from inventory at checkout.
Visual identification technologies are becoming more common. People use these technologies everyday in video games that allow users to mimic dance moves and swing virtual tennis rackets.
"There are millions of devices in the hands of retailers and consumers across the United States today that can take a picture of an item — an apple or a detergent bottle, for example — send that picture to a server and then let software automatically identify the item," says Justin Patton, managing director of the RFID Center. "The process is as simple as scanning a barcode, but it works from all angles and can capture many items at once. An image is enough, and the hardware is cheap."
The university centers will collaborate on research to make visual identification technologies reliable for business applications, including identifying items missing from store shelves, speeding up self-checkout and locating items in distribution centers. The researchers will then work with global standards organizations to develop open standards that will allow for broad adoption.
"What is needed is a common digital data standard for shared items that can be used across platforms and amongst retailers," says Jackson Cothren, associate professor of geosciences and director of CAST. "You can buy a bag of chips at a concert in Germany or at a gas station in New Mexico, and the standard VIT traits of that item are shared. This will have a major impact on the retail industry in the next three years."
The RFID Research Center serves a large network of retailers and suppliers and has broad knowledge of store processes. The center has been successful at integrating new technology at large retailers such as Walmart, Dillard's, JC Penney, and Bloomingdale's.
CAST has developed two-dimensional and three-dimensional analytics for geospatial location and mapping and shape analysis of natural and manmade objects. The center has developed digital standards for collecting and archiving of these data.
Together, the RFID Research Center and CAST have identified the most effective emerging technologies, all of which can be integrated with open databases to produce simple, fast and inexpensive visual identification technologies.
Preliminary research began in 2010 with early prototypes of Microsoft Kinect systems that can identify grocery items. The centers are now conducting studies on deep scans of actual store environments and within laboratory space at the RFID Research Center. The researchers have created 3-D models of the products and store environments comprised of billions of data points. They have begun to identify promising data that can be used to create a standard library of object information. Initial results will be issued in the fall of 2013.
No entries found