Microsoft's New App Can Read Your Emotions
So how were you really feeling in that selfie?
Microsoft has unveiled a new app that can tell you exactly how you were feeling when the shutter went off. This demo is from Microsoft’s Project Oxford, which has previously offered a variety of photo recognition tools including the popular age detection engine. However, this new app takes it to the next level.
The app breaks down your face into emotional constituents and comes up with a score card relating to a set number of emotions including anger, contempt, disgust, fear, happiness, neutral, sadness and surprise. While it obviously has its flaws and drawbacks, the demo works off a learning process that builds its predictions over time. So the more photos it reads, the more it will improve.
To try this out, we plugged in a photo of Kanye West and Kim Kardashian. It’s surprisingly fast and while we can’t really say whether it’s accurate, it does do a good job of inferring possible emotions. For Kanye, the machine deemed his emotion to be mostly neutral, with a bit of contempt, happiness and sadness. Kardashian on the other hand, scored mostly in happiness, with a bit of contempt, neutrality and surprise. While we’re sure a machine can’t (yet) detect notions of irony or other more subtle emotions gleaned from body language and experience, it sure is an interesting advancement. Let us know in the comments what you think, and you can try it out yourself here.
