Google Im2Calories App: Counting Calories from a picture
We don’t know about you, but we’re becoming a little social media self-conscious these days. We spend hours on end composing tweets and we’re forever detagging photos on Facebook. But more than anything else, we’re getting self-conscious about showing the world our meals on Instagram. Well, we’re tired of the ridicule and now we have a reason to carry on regardless thanks to Google’s Im2Calories app!
The idea is simple – the Im2Calories uses ‘deep learning’ algorithms to calculate how many calories are in a meal from nothing more than an Instagram-sized image. While the idea is simple, the technology is obviously a little more head-scratching.
According to Google, Im2Calories uses image-recognition, in addition to widely available caloric information, to understand what the food is that you’ve got on your plate and tell you the estimated sum of calories you’re about to eat.
The ‘deep learning’ bit is the really important bit here. Rather than relying on loads and loads of initial data to start off with, Im2Calories will improve its accuracy through recognising its own mistakes over time.
Minimal input will be needed by users initially if/when the app recognises your asparagus as lumps of cheese but the creator, Kevin Murphy, is quoted as saying he’ll be happy if he gets “the calories off by 20%.” This is because Google can collate the data and average out predictions.
So while Im2Calories won’t ever be 100% accurate, its use isn’t completely redundant. Many of us Janes aren’t as meticulous with our calorie counts and we’d like people to believe anyway! We say, whatever helps. And let’s be honest, it’s just a very cool idea!
Google has filed patents for the technology already but annoyingly there’s very few details on when Im2Calories will be available just yet. So just watch this space!