At the conference Rework Deep Learning Summit researcher Google Kevin Murphy presented an application that is able to calculate the caloric content of meals on his photos. In the future to obtain such information will simply photograph your food on camera phones.
Service called Im2Calories uses a set of algorithms for deep learning DeepMind bought early last year. While it exists only as a research project. Application using artificial intelligence technology, analyzes the image of the products both in absolute terms and relative to the plates. It is not necessary that the photo was high resolution — enough picture from Instagram.
Im2Calories can also determine the type of seasoning. At the presentation of the application to recognise in the photo two eggs, two pancakes and three strips of bacon.
At the moment the developers have a problem with the accuracy of counting calories. Now the program determines the number of calories in a 20% chance. To do this, the creators Im2Calories use the values that are indicated in the documents of various products. The application must be a “diary” and a tool that facilitates the documentation in the fitness developments eaten for a day of food.
As noted by The Verge, in 2011, researchers conducted a study according to which counting calories has such a significant value, because the value also has quality, not quantity of food. In 2010, the head of the organization Weight Watchers David Kirchhoff said that counting calories is not of great importance for those who want to lose weight.