When you think of smiling, you usually equate it with happiness. While that’s often the case, it isn’t always; people have been known to smile for other reasons as well. Perhaps the second most common reason to smile is to fake happiness, with smiling out of frustration coming in third. As a person, you’re usually able to suss out the meaning behind a smile by looking at it in context of the moment in which it happened, but given a single image, your accuracy goes way down. New smile-tech from MIT, however, excels at figuring out the story behind the snapshots.
As it turns out, the physical characteristics of a smile tend to be different depending on the stimulus. Involuntary smiles, for instance, are subtly different from voluntary ones. Likewise, ones that well up from total delight are different from those that inexplicably creep out while you’re trying to cope with frustration. When presented with single frames, MIT’s smile algorithm can recognize frustrated smiles with 90% accuracy whereas regular humans score worse than the percentages they’d get with random chance.
Okay, but so what? It’s not often that a person has to tell whether someone is happy or frustrated with just one out of context photo. True, but kids with autism, though they have a lot of context, still have a very difficult time picking up on context clues, especially differentiating between different kinds of smiles that could mean drastically different things. By teaching autistic kids to use the same variables the algorithm uses, it’d be possible to help them compensate for their lack of intuition on that front.
Check out the video below to learn more, and make sure to get your fake, polite smiles in while you can; it might not be long before someone — or something — can call you out on them.
(via The Next Web)
