Get a prediction of how other people perceive your face | Hacker News Reader

Get a prediction of how other people perceive your face

face.cbs.dtu.dk

151 pointsdhsb37 comments

Open articleView on HN

"Error, code UL1JOBHASH: parsererror"

I knew I wasn't the best looking guy but this seems more harsh than I anticipated.

I got that too.

Same for me, and I'm beautiful :P

I suspect their server has succumb to the Hacker News effect

haha

Dear all,

Unfortunately, we are currently experiencing technical difficulties with the Face It! project website. This is most likely due to the rapid increase in web traffic, after the project was submitted to Hacker News, today April 8 2015. We greatly appreciate the increased interest our project has generated, and are currently working at full throttle to solve the current server issue. Hopefully, the website will run at full functionality again very shortly. We apologize for any inconvenience this may have caused. We will, on Hacker News and the Face It! project website, keep you updated on the status of the project web service as a solution progresses.

Please stay tuned...

Dear all,

We are happy to tell you, that the web service is now functional again. It can still be visited at the original URL at http://face.cbs.dtu.dk/. We would like to hand out a big thanks to intrbiz from the IRC channel #suse on the freenode IRC network, for his great support in helping us getting our server properly recovered again. We are sorry for the inconvenience caused in the mean time, but hope you will enjoy the final work.

The first thing I thought when it didn't load was "maybe this is a study to see how persistent people are to find out if they are attractive?" It's probably just the HN effect though.

This is a bit sensitive to cultural influences. For instance, male facial characteristics have a clear regional variance, thus biasing the result in a particular population. I can tell from experience, I'm a male pushing over thirty with ridiculously androgynous features. I'm tagged as a boy by local population, and when I visit UK, I'm tagged as a 'ugly woman' :)

You will be fine, because you can tell the program that you're male.

However, the study has been using 244 Danes... I think you'll be right that cultural influences will be a defining factor here. I don't think the majority of people looks like a Dane. :-)

I hope they don't tell me that I look like a great dane.

"male facial characteristics have a clear regional variance"

?????

Link to the actual paper if anyone is interested:

http://www.ncbi.nlm.nih.gov/pubmed/25233221

There've been times where I would just upload a picture of myself to the internet. For fun.

As noted in the first paragraph, a tool such as this will be useful to "other cultures" only after a panel comprised of people of the target culture have calibrated the tool.

> No hair obscuring the face.

Welp I guess my beard disqualifies me. I wonder how many HNers have the same problem?

I'm getting a parser error, that's seems to explain why.

I'd be really interested to see someone take the research they did here, and extend the survey to various different cultures. Right now, this model will give you a prediction of how a sample of 244 danes will perceive your face, which may be at least somewhat representative of the danish population, but likely not too much beyond that.

It should be easy enough to survey people from different cultures and countries, and feed those surveys into the model, to get a more accurate representation. You could even take that, and have a global model, as well as different models based on cultures, to get things like "This is how you will be perceived by most people in the southern united states".

This has got me humming 'there is nothing like a Dane' - which is not quite a song from the musical South Pacific.

The link is taking time to load. Looks like everyone wants to know what others think about them. Expected.

Has this actually worked for anyone?? Not one commenter seems to have got it to work.

I've had nothing but an error ("Error, code UL1JOBHASH: parsererror") for the last two hours.

> This study is based on results from 244 Danish participants

That's a small learning set. I suppose now that it's on HN front page, they will get significant amount of data to tune their model across cultures and nationalities.

Hypothetically, wouldn't it be easier for them to train their model using the portfolio images of public figures like actors/politicians etc for whom I'm sure there would be tremendous amount of subjective description online.

We will get our chance to play with this -- it will just take a while. I emailed the researchers about site snarls, error messages, etc., and they wrote back that all our attention

:... is definitely welcome, even though it’s very unexpected and therefore overloaded the server! Ulrik Plesner, who developed the webserver, is working at the moment to get it up and running again..I think it'll be back over the next 1-2 days if all goes well :) "

I have a prediction based on 50 years of empirical, hard-won data...

Old and Unattractive.

:)p

Down for me. A text only version is avalibal from Google cache: http://webcache.googleusercontent.com/search?q=cache:7YJ7VTW... , but not sure that gives to much sens.

Would be nice to see some examples of each extreme, for each dimension.

There: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4169442/figure/p...

It looks like that in Danish society, being darker may be correlated with perceptions of dominance.

I found a random image of a human face and tried to upload it. Stuck at "Image queued for upload..." for minutes. Then came a page asking me to "annotate" the image by adjusting the auto-positioned point markers on my image according to this image[0]. Finally I saw this message: "An error has occured. Please try again." Is the server too busy right now?

[0]: http://face.cbs.dtu.dk/graph/annoEx.png

Mirror: https://archive.is/YAaex/5d06f6875b232350ff6954739af37c44a3d...

Site is not opening Seems like HN effect has freezed their server.

Perhaps too many ugly people are using the server and it refused to continue analyzing pictures!

Doesn't test anything for me for some reason. Not sure if I'm doing something wrong but it doesn't upload or show any 'Annotation points'.

Edit: Just saw this : "We are currently encountering server errrors, resulting in the service being unable to run. We are working to get this up and running as soon as possible. Please stay tuned... "

In addition to what the first paragraph says, I'll add that given the strong cultural bias there can be on such thing, I wouldn't put too much trust into such a tool except if it could be calibrated by a panel of the "other people" you are targeting.

After the error the button isn't even working. I think there's someone just trying to collect photos...

commence the HN slashdot.

Reply on news.ycombinator.com