ChickenRiceNot all food is cooked equal. A cheeseburger cooked at home versus one bought at a fastfood joint can be very different in nutritional value due to the choice of ingredients used during the preparation process, which is why sometimes counting calories might not necessarily be very accurate, but it is still a good way of getting a rough idea of how much you are consuming.

However to help make the process of calorie counting more efficient and accurate, it has recently been revealed that Google has been working on a project that taps into artificial intelligence to help analyze photos of food and spit back out the amount of calories there are in that meal.

This was revealed during the Rework Deep Learning Summit where Google research scientist Kevin Murphy first unveiled the project. Dubbed “Im2Calories”, it relies on a camera to analyze the photo and based on the demonstration that was put on then, it seems that the system is pretty adept at recognizing the different elements on a plate of food. It also gauged the size of each piece of food in relation to the plate, and also took into consideration any condiments that might have been used.

So how accurate is Im2Calories? At the moment it is unclear as to how close or far the system is from its readings, but Murphy states that Im2Calories is a system which will improve itself through use overtime. “To me it’s obvious that people really want this and this is really useful. 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.”

At the moment there are apps like MyFitnessPal which has a huge database of different types of food, portion sizes, and etc. which users will have to enter in manually, but if Im2Calories could speed up the process through automation, we reckon it would have an edge over the competition. Unfortunately there is no word on if and when it will actually be realized into an actual product.

Filed in General. Read more about and .