Can emotions be quantified? … Microsoft says yes!
Sebah Al-AliWhat’s worse than not being able to sleep at night? Sleeping intermittently, because an inconsiderate neighbor’s tiny yet very loud dog is kept outside, barking all night…. with breaks in between. You’d think it’s over at 1 AM, and fight yourself to go back to sleep…. But, 53 minutes later Mr. tiny dog is feeling lonely again… and, again and again. Yes. I was this close to ringing their doorbell at 2:30 AM, with a grumpy face and a piece of my mind. I just was too tired to leave my bed.
In addition to a ton of regret that I didn’t actually ruin my terrible neighbor’s night like he ruined mine, I clearly have a mix of negative feelings towards this terrible experience: anger, frustration, exhaustion, hatred probably, and a few other feelings. A lot of feelings that I have experienced first hand, but I don’t think I can give weight to each one of them. Can you?
Well, Microsoft thinks it can! Microsoft has developed an emotion recognition algorithm that analyses facial expressions to identify emotions on faces, by numbers! After you upload a picture, Microsoft emotion detection engine can detect eight basic facial emotions: anger, contempt, fear, disgust, happiness, neutral, sadness, and surprise.
Give it a try yourself: https://www.projectoxford.ai/demo/Emotion
Interesting, right? I was wondering how helpful this can be to teach our students vocabulary about emotions. We can probably ask them to draw a face expressing one of these emotions and then see how accurate it is using Microsoft’s emotion engine. The one that gets the highest score wins? If they’re male students you can probably ask them to make these faces themselves and see how well they can show their feelings.
It’ll be a lot of fun! Don’t you think so?