Emotion Recognition Technology on the iPhone?

iPhone users: Blue Finger, Inc., a South Korean development company, has released a new iPhone app that reads facial expressions from photos (as well as in real time, if you purchase the full version.) The app scans the faces and gives you one of 12 answers: happiness, excitement, anger, joy, worry, loneliness, tiredness, sadness, “poker-face,” surprise, boredom, and “no-idea.”

Although technology capable of emotion recognition has been in development, the scientific validity of the answers the app comes up with is questionable. For one, feelings such as fatigue, loneliness, and boredom are not expressed the same way universally, so it would not be possible to develop an algorithm that correctly detects such feelings for every individual person. While happiness and anger are expressed universally  all over the world, it is doubtful that a tiny iPhone app is powerful enough to detect these emotions in real time.

We played around with the app in the office, taking photos of ourselves making various expressions and seeing what Mind Scan would come up with. It felt as though the answers it gave us were at random.

For example, an obviously disgusted expression was read as one of surprise, and a big smile was read as tiredness. The app doesn’t seem to cater to those who would care too much about the science behind facial expressions, but the readings were way too far-fetched to even be entertaining.

Looking at the reviews from the iTunes App Store, it seems as though a good amount of people have had fun with the app, and it has a 3 out of 5 star rating. However, most people who gave Mind Scan a positive rating stated that it is definitely a novelty app and should not be taken seriously.

Are you an iPhone user? Do you think you will download the Mind Scan Camera App?

For more information, as well as a video about the app, click here.

3 responses to “Emotion Recognition Technology on the iPhone?”

  1. Russ Conte says:

    I use the iPod Touch (the iPhone minus AT&T) and I’ll pass on this app. There are several similar apps in the app store, and none of them are genuine – they would not pass any test of reliability that I can imagine. It’s very interesting, and I’m sure apps will mature as time goes on (and as developers get feedback) but this is just a game, I would not take it (or any similar apps) seriously. At least not yet 🙂

  2. Markus says:

    I would never use this app. As already pointed out the reliability and validity seems horrible.

    I assume that that intention is not to take it seriously just for fun.

    The other thing i have against all computerized facial reading is that it might impair the natural way of reading facial expression even more. My opinion is that people get more “lazy” with technology and put to much trust in it.

    Reading faces and emotion takes understanding of behaviour and context. Wich cant be programed in to an app. Its an human interpesonal “art” that cant be put into mathematical equations in a software.

  3. Steve says:

    While this app doesn’t have the ability to even come close to being accurate, I have no doubt that a reliable program will be developed in the near future. It will need more processing power than a phone has, but it will happen.

    I just watched NOVA’s episode on the IBM “Watson” computer that recently won the game show “Jeopardy”.

    It detailed how difficult it is for a machine to understand language and how “Watson” was programed to “learn” how to answer questions correctly. It was very impressive, maybe scary.

    Compared to that, programming a computer to learn how to read body language, facial expressions, gestures, and even seeing micro expressions, wouldn’t take much more effort.

    I agree with Markus about people getting “lazy” with technology and putting too much trust in it at times. But technology is advancing exponentially. It’s standing next to the “creepy line” and looking where it will take its next step.

Leave a Reply

Your email address will not be published. Required fields are marked *

Copyright © Humintell 2009-2017