Skip to main content

Shopping tool for the visually impaired

Visually impaired people who want to go grocery shopping independently may eventually benefit from a cell-phone sized "reading" and "seeing" device being developed at UCSD. Similar hand-held gadgets could provide the visually impaired with more autonomy in places like airports and bookstores if additional research is funded and prototypes perform as expected, explained Serge Belongie, the computer science and engineering professor who leads the research team. To use the handheld device, now called a "MoZi box," users will first create image-and barcode-based shopping lists on their home computers. When someone is ready to shop, he or she will download the list to the MoZi box, which is also outfitted with a camera and other navigational tools to help users get around and avoid obstacles. Once at the store, the MoZi box will "read" aisle signs. If Tide laundry detergent were on a grocery list, the MoZi box would alert the user when he or she walked past an aisle labeled "laundry, household cleaning, dish detergent." Next, based on crunched and stretched versions of the images of Tide products downloaded at home, the MoZi box will scan the aisle for Tide boxes using an image recognition system. Once a potential match is located, the MoZi box will scan the barcode for verification. It will also help the user find the checkout and, when paying, differentiate between a five and a fifty. The project, part of Belongie's ongoing studies into mobile computer vision, includes Jacobs School students, Calit2 researchers, and collaborators from the University of Kentucky.

Calit2 researcher John Miller ( Ph.D. '03, electrical engineering), who is blind, tests the prototype MoZi box shopping tool at a Mira Mesa Vons grocery store near the UCSD campus.
Calit2 researcher John Miller ( Ph.D. '03, electrical engineering), who is blind, tests the prototype MoZi box shopping tool at a Mira Mesa Vons grocery store near the UCSD campus.