I wonder if there is any freely available database of facial pictures linked to specific emotions. It would be pretty simple for someone to make a “guess that emotion” web game.
ETA: I took the test myself and scored below normal:
Your score: 20 A typical score is in the range 22-30. If you scored over 30, you are very accurate at decoding a person’s facial expressions around their eyes. A score under 22 indicates you find this quite difficult
I did, indeed, find the test extremely difficult. I usually look at the mouth more than the eyes when trying to read faces...
I wonder whether the ‘right answers’ are what the subject of the photograph was actually feeling, what an expert intended the photograph to represent, or what most people respond.
Some of the pictures of women look like they could be movie stars or models. The “fantasizing” eyes look more like what you’ll see on the cover of a fashion magazine than like anything I’ve ever seen in real life.
Holy crap, I just scored 33. I guess all that hard, bitter experience paid off.
Edit: I am a bit surprised to read about how hard other people found it. The options tended to have quite a predictable multiple-choice structure: two obviously implausible answers and two plausible ones. #21, for example, doesn’t have the tension to be either confused or panicked. Embarrassment is plausible, but she doesn’t look coy enough.
For the record, I am nowhere near this adept at spotting women engaged in the act of fantasising in real life. Multiple choice faces would be a great, if unrealistic, help.
That’s an interesting test! I scored 28, and for most of the questions I failed, I find that the correct answers make much more sense than my initial guesses when I look back at the pictures. I find #29 the most cryptic.
Look at the direction of the gaze and the lack of eye contact. When someone is facing you but looking away, they are either thinking of something else or being avoidant. If they are being avoidant and looking down, then they are generally being submissive, possibly because they are lying.
Score of 30 btw. I had the same experience where I was able to recontextualize the pictures. It would make a great application if it had like 100 pictures.
Edit: Just a thought, would it be useful to people to have an article or something describing specific things to look for, like I have above? Sort of like an emotional taxonomy? It seems like it’d be an interesting project.
I think I used a combination of gaze/eye contact and eye shape, eg someone with very wide open eyes is probably scared or surprised, someone with narrowed eyes is concentrating/suspicious, etc.
Incidentally, my own score was 29 and none of my errors were completely implausible. This also goes a long way to explaining for me why it bothers me so much when people wear sunglasses.
would it be useful to people to have an article or something describing specific things to look for, like I have above
I think it would be great, but might be hard to include sufficient detail to be useful to people who have difficulties. Preoccupied and guilty both involve lack of eye contact but how do you describe the other differences?
Well for those two, I think there are two possible differences. Guilty people are going to be looking down more and possibly have there eyes closed a bit more.
Possible attributes off the top of my head:
eye contact
general eye direction
direction of the face
open/closed
eyebrows
head tilt (look at number 21for a great example)
wrinkled/furrowed face
Are there any more? That’s definitely enough to categorize most emotions.
When you think about it, it’s quite telling about just how expressive our eyes are when this test expects people coming out substantially below average to still pick correctly twice as frequently as if they were just guessing.
35.
It felt hard, but apparently I’m better at this that I’d have thought. Really surprising how subtle this is—modeling expression recognition would be a hell of a challenge.
I scored 31; my first reaction on seeing “A typical score is in the range 22-30” was that it they deliberately gave a lower range, so as to give a better feeling to the user (like those “Wow you matched the ‘winner’ profile by 98%! You should buy our product!” online tests).
But looking at comments here, and at the distribution given in the paper you linked, I guess I did get a decent score. Maybe my methodology worked better (hide the words, just look at the picture, describe the expression to myself, then pick the word that seems the closest); or all those years drawing stick-figure comics with exaggerated expressions (and iterating a lot on how to draw them right) did pay off.
I mean, I certainly don’t consider myself above average in social skills :D
24 here. Higher than I thought I’d get. I was distracted by how my peripheral vision was filled with the other faces. And in real life, you do get the whole face, which gives context to the eyes.
I got a 23, myself; considering that I’m a diagnosed Aspie, that’s not too bad, I suppose. I can pretend to be normal fairly well, anyway; it’s mostly the stuff about getting stuck in routines that trips me up nowadays.
How good is your monitor? Last night I solved the test on a laptop whose greyscale color resolution isn’t too great, and I also didn’t notice that my browser had downsized the images. Now when I look at it on a better monitor, almost all the images seem much clearer and easier to interpret.
I wonder if there is any freely available database of facial pictures linked to specific emotions. It would be pretty simple for someone to make a “guess that emotion” web game.
Actually, something like this exists:
Reading the Mind in the Eyes
Apparently, people with high-functioning autism or Aspergers do much worse than control subjects.
ETA: I took the test myself and scored below normal:
I did, indeed, find the test extremely difficult. I usually look at the mouth more than the eyes when trying to read faces...
I wonder whether the ‘right answers’ are what the subject of the photograph was actually feeling, what an expert intended the photograph to represent, or what most people respond.
I wonder that too.
Some of the pictures of women look like they could be movie stars or models. The “fantasizing” eyes look more like what you’ll see on the cover of a fashion magazine than like anything I’ve ever seen in real life.
Holy crap, I just scored 33. I guess all that hard, bitter experience paid off.
Edit: I am a bit surprised to read about how hard other people found it. The options tended to have quite a predictable multiple-choice structure: two obviously implausible answers and two plausible ones. #21, for example, doesn’t have the tension to be either confused or panicked. Embarrassment is plausible, but she doesn’t look coy enough.
For the record, I am nowhere near this adept at spotting women engaged in the act of fantasising in real life. Multiple choice faces would be a great, if unrealistic, help.
30 here. And I got all the fantasizing/desire ones.
That’s an interesting test! I scored 28, and for most of the questions I failed, I find that the correct answers make much more sense than my initial guesses when I look back at the pictures. I find #29 the most cryptic.
Look at the direction of the gaze and the lack of eye contact. When someone is facing you but looking away, they are either thinking of something else or being avoidant. If they are being avoidant and looking down, then they are generally being submissive, possibly because they are lying.
Score of 30 btw. I had the same experience where I was able to recontextualize the pictures. It would make a great application if it had like 100 pictures.
Edit: Just a thought, would it be useful to people to have an article or something describing specific things to look for, like I have above? Sort of like an emotional taxonomy? It seems like it’d be an interesting project.
I think I used a combination of gaze/eye contact and eye shape, eg someone with very wide open eyes is probably scared or surprised, someone with narrowed eyes is concentrating/suspicious, etc.
Incidentally, my own score was 29 and none of my errors were completely implausible. This also goes a long way to explaining for me why it bothers me so much when people wear sunglasses.
I think it would be great, but might be hard to include sufficient detail to be useful to people who have difficulties. Preoccupied and guilty both involve lack of eye contact but how do you describe the other differences?
Well for those two, I think there are two possible differences. Guilty people are going to be looking down more and possibly have there eyes closed a bit more.
Possible attributes off the top of my head:
eye contact
general eye direction
direction of the face
open/closed
eyebrows
head tilt (look at number 21for a great example)
wrinkled/furrowed face
Are there any more? That’s definitely enough to categorize most emotions.
A few important ones
pupil dilation
narrowing of eyelids
wrinkles in the bridge of the nose
wrinkles in muscles around the eyes
symmetry of movement
When you think about it, it’s quite telling about just how expressive our eyes are when this test expects people coming out substantially below average to still pick correctly twice as frequently as if they were just guessing.
35. It felt hard, but apparently I’m better at this that I’d have thought. Really surprising how subtle this is—modeling expression recognition would be a hell of a challenge.
I scored 31; my first reaction on seeing “A typical score is in the range 22-30” was that it they deliberately gave a lower range, so as to give a better feeling to the user (like those “Wow you matched the ‘winner’ profile by 98%! You should buy our product!” online tests).
But looking at comments here, and at the distribution given in the paper you linked, I guess I did get a decent score. Maybe my methodology worked better (hide the words, just look at the picture, describe the expression to myself, then pick the word that seems the closest); or all those years drawing stick-figure comics with exaggerated expressions (and iterating a lot on how to draw them right) did pay off.
I mean, I certainly don’t consider myself above average in social skills :D
“Social skills” is one of those terms that’s in dire need of explosion.
I was surprised to get 27; apparently I’m better at it than I thought.
One of the main hints I used was “eyes-on” vs. “eyes-off”; whether they were looking towards me or not.
The other main one was the tension in the face.
I found this extremely difficult too! Scored 21, and frankly, I was surprised to get that many.
25.
I found the test surprisingly fun.
24 here. Higher than I thought I’d get. I was distracted by how my peripheral vision was filled with the other faces. And in real life, you do get the whole face, which gives context to the eyes.
I got a 23, myself; considering that I’m a diagnosed Aspie, that’s not too bad, I suppose. I can pretend to be normal fairly well, anyway; it’s mostly the stuff about getting stuck in routines that trips me up nowadays.
Scored 31. Wasn’t an easy test. The reproductions are terrible.
How good is your monitor? Last night I solved the test on a laptop whose greyscale color resolution isn’t too great, and I also didn’t notice that my browser had downsized the images. Now when I look at it on a better monitor, almost all the images seem much clearer and easier to interpret.
I got 31. Apparently I’m bad at autism.
It’s a very interesting test though. The meanings of eye expressions are less context dependent than I’d have thought.
The wikipedia entry on microexpressions has links to both free and commercial software.
Great idea, it could have video as well. This is something that if no one has done it before could get funding.