Calories counted using photos, crowds

Americans spend $40 billion a year on dieting advice and self-help books but the first step to healthy eating is knowing how much you eat, U.S. researchers say. To make keeping a food diary easier, computer scientists at the Harvard School of Engineering and Applied Sciences devised a tool that lets people snap a photo of a meal and let crowd-sourcing -- an invited crowd of people -- determine the caloric content. PlateMate's calorie estimates proved, in tests, to be just as accurate as those of trained nutritionists, and more accurate than the user's own logs, the researchers found. Jon Noronha, who co-developed PlateMate as an undergraduate at Harvard with Eric Hysen, said PlateMate's advantage is that it allows the user to consult impartial observers quickly, without having to pay for the advice and supervision of an expert nutritionist. PlateMate works in coordination with Amazon Mechanical Turk, crowd workers who receive a few cents to distinguish between foods in the photo, identify what they are and estimate quantities. The nutrition totals for the meal are then automatically calculated, the researchers said. The findings were presented at the 24th ACM Symposium on User Interface Software and Technology in Santa Barbara, Calif.