Manually counting calories may soon become a thing of the past if Google has anything to do with it. Speaking at the Rework Deep Learning Summit in Boston, Google unveiled a project called Im2Calories which uses deep learning methods to calculate the calories of food from just about any photo.
The project was introduced by Kevin Murphy, a Google research scientist. Im2Calories is able to scan a photo and use deep learning to determine not only the types of food in it – for example two eggs, two pancakes and three strips of bacon, but also the approximate size in relation to dishes and condiments also in the photo. The process allows for users to correct the app should it misread a certain food, and the user input then allows Im2Calories to learn for better analysis in the future.
“We semi-automate. 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.
As more data is analysed, performance increases and the process becomes more accurate and requires less input from users. Don’t expect an app based on this technology any time soon as Google only recently applied for a patent on it and Murphy was tight lipped as to when it might be made publicly available. Google is only just beginning though as Murphy indicates that counting calories is just the start.
“If we can do this for food, that’s just the killer app. Suppose we did street scene analysis. We don’t want to just say there are cars in this intersection. That’s boring. We want to do things like localize cars, count the cars, get attributes of the cars, which way are they facing. Then we can do things like traffic scene analysis, predict where the most likely parking spot is. And since this is all learned from data, the technology is the same, you just change the data.”
What do you think about Google’s latest push to drive deep learning for tasks like calorie counting? What other uses do you hope Google focuses on? Let us know in the comments below, or on Google+, Twitter, or Facebook.Source: Popular Science