But as any calorie app user knows, sometimes selecting the right food can yield wildly different results from what is actually being consumed. Google’s new technology aims to take the guesswork out of calorie counting with software that estimates how many calories appear in a photo using artificial intelligence.
Im2Calories was unveiled by Google research scientist Kevin Murphy at Boston's Rework Deep Learning Summit. In the demonstration the technology was able to accurately count the calories on a breakfast plate filled with eggs, toast and bacon, reports Popular Science.
No high-resolution picture is needed, the software works by recognizing other similar images on the web through visual analysis of pattern recognition and connects it to available caloric data. Murphy says that it often misreads dishes but the app contains a dropdown menu to adjust cooking styles-- for example if a fattier fried egg is confused for a poached one.
“If it only works 30 percent of the time, it's enough that people will start using it, we'll collect data, and it'll get better over time,” said Murphy during the demonstration.
Though Google recently filed a parent for Im2Calories, it might be awhile, if ever, for the product to get to the mass market.
“To me it's obvious that people really want this and this is really useful,” Murphy told the conference. “Ok fine, maybe we get the calories off by 20 percent. It doesn't matter. We're going to average over a week or a month or a year. And now we can start to potentially join information from multiple people and start to do population level statistics. I have colleagues in epidemiology and public health, and they really want this stuff.”
It may be a welcome tool for many but twenty percent can make a big difference for serious dieters. Maybe Im2Calories will be about as accurate as Microsoft’s dubious app for figuring out age from a picture.
