Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Are 625 Pixels Enough To Identify Sex?

timothy posted more than 3 years ago | from the when-tweet-analysis-fails dept.

AI 143

mikejuk writes "A Spanish research team have patented a video camera and algorithm that can tell the difference between males and females based on just a 25x25 pixel image. This means that there is enough information in such low resolution images to do the job! They also demonstrate that an old AI method, linear discriminant analysis, is as good and sometimes better than more trendy methods such as Support Vector Machines..."

cancel ×

143 comments

Sorry! There are no comments related to the filter you selected.

It's small (0)

Anonymous Coward | more than 3 years ago | (#35844402)

But it's not that small. At least that's what my girlfriend tells me.

I am pretty sure that I... (0)

Anonymous Coward | more than 3 years ago | (#35844406)

can do it with fewer pixels.

Re:I am pretty sure that I... (2, Funny)

Anonymous Coward | more than 3 years ago | (#35844456)

Here you go:

Male: |
Female: O

Re:I am pretty sure that I... (0)

Anonymous Coward | more than 3 years ago | (#35844724)

But can it detect Rosie O'Donnell or Chris-chan's gender?

Re:I am pretty sure that I... (0)

Anonymous Coward | more than 3 years ago | (#35845618)

or hilary clinton's?
or nancy pelosi's?

Re:I am pretty sure that I... (4, Funny)

Joce640k (829181) | more than 3 years ago | (#35846266)

...so can anybody from the old BBS days.

Re:I am pretty sure that I... (1)

riT-k0MA (1653217) | more than 3 years ago | (#35846368)

My kingdom for a mod point...

Depends... (5, Funny)

mmaddox (155681) | more than 3 years ago | (#35844416)

...on what it's an image OF.

Am I the only person imagining genitalia icons?

Re:Depends... (4, Funny)

93 Escort Wagon (326346) | more than 3 years ago | (#35844496)

Am I the only person imagining genitalia icons?

Yes.

Re:Depends Are For Old Farts (-1)

Anonymous Coward | more than 3 years ago | (#35844580)

You are a nigger.

Re:Depends Are For Old Farts (0)

Anonymous Coward | more than 3 years ago | (#35845676)

I done my robe and wizard hat...

Re:Depends... (0)

Anonymous Coward | more than 3 years ago | (#35844626)

Am I the only person imagining genitalia icons?

Tell me about your mother.

Re:Depends... (1)

davester666 (731373) | more than 3 years ago | (#35844630)

Somebody notify the IOC!

No need for any fancy genetic testing anymore.

Re:Depends... (-1)

Anonymous Coward | more than 3 years ago | (#35844978)

Nope, you're not... (.Y.)

Re:Depends... (1)

thatskinnyguy (1129515) | more than 3 years ago | (#35845304)

This one time, I accidentally a whole word.

Re:Depends... (0)

Anonymous Coward | more than 3 years ago | (#35845754)

This is just promoting i-programmer.info again. I believe this is like the 5th article submitted that links back to that crappy wordpress website. I normally don't say things that aren't related to the top thread but seriously this is getting serious when crappy sites with unverifiable content gets posted here. Oh wait, I must be new here etc etc...

Hm? (5, Funny)

Auto_Lykos (1620681) | more than 3 years ago | (#35844420)

"The also demonstrates that an old AI method, linear discriminant analysis, and demonstrates that it is as good and sometimes better than more trendy mehods such as Support Vector Machines"

I think the summary accidentally forgot the

Re:Hm? (4, Funny)

zill (1690130) | more than 3 years ago | (#35844432)

I think you forgot the

Re:Hm? (0)

diesel66 (254283) | more than 3 years ago | (#35844486)

That's pretty funny. Not only did you forget it, but so did the

Re:forgot (1, Funny)

TaoPhoenix (980487) | more than 3 years ago | (#35844718)

(

(xkcd ftw again!)

Re:forgot (2)

pushing-robot (1037830) | more than 3 years ago | (#35845410)

)

Oh, what sad times are these when passing ruffians can open parenthesis at will on Internet forums.

Re:Hm? (1)

DEmmons (1538383) | more than 3 years ago | (#35844720)

The is a problem, but I'm pretty sure the mehods they use on images in CSI could 'enhance' the missing data back into that sentence, which also demonstrates that a good algorithm.

Re:Hm? (1)

AK Marc (707885) | more than 3 years ago | (#35845024)

All digital cameras need is a 25x25 CMOS and the CSI software to get better results than the Hubble. How long does it take to get a 30 megapixel image from the 25x25 with CSI processing?

Re:Hm? (1)

Anonymous Coward | more than 3 years ago | (#35845238)

How long does it take to get a 30 megapixel image from the 25x25 with CSI processing?

CSI Software runs at... the Speed of Plot [tvtropes.org] .

*sunglasses* YEEAAAHHH!

Re:Hm? (0)

Anonymous Coward | more than 3 years ago | (#35845256)

one two parter

Re:Hm? (1)

zippthorne (748122) | more than 3 years ago | (#35845774)

Pfft. With CSI processing, you can take a picture of a spoon with a one-pixel camera, and get your detailed, holographic image of the moon from the reflection off the spoon....

Footnote (4, Funny)

RenHoek (101570) | more than 3 years ago | (#35844428)

Works on a 25x25 pixel image*

(* Pixels need to be a shade of pink)

not believable (1)

Anonymous Coward | more than 3 years ago | (#35844434)

I thought this story might be believable until I looked at the page. I'm not 100% sure what gender the 2nd row, 4th from the left person is and by the way, I'm a human. So I think the rest of the title to this story is "an arbitrarily acceptable percentage of the time so oh just publish it, it sounds neat"

Re:not believable (0)

Anonymous Coward | more than 3 years ago | (#35844478)

I thought this story might be believable until I looked at the page. I'm not 100% sure what gender the 2nd row, 4th from the left person is and by the way, I'm a human.

Where do we file a bug report for the Slashdot chat bots? This one just read the article...

Re:not believable (1)

PyroMosh (287149) | more than 3 years ago | (#35844584)

The images shown were ones the software had trouble classifying (It got them wrong). The top row are male, the bottom row are female. This is all explained in the caption.

Interestingly the article does not make any mention of error rates. And I couldn't find anything easily on either site it links to.

Re:not believable (1)

osu-neko (2604) | more than 3 years ago | (#35845088)

... This is all explained in the caption.

And I'm laughing my ass off. I knew /. "readers" rarely RTFA, but I always assumed they were just too lazy to follow the links. I never realized they actually click the links, but only to look at the pretty pictures. xD

Re:not believable (1)

Anonymous Coward | more than 3 years ago | (#35844602)

Tim's keep pulling one bullshit story after another. Pathetic even for slashdot "standard".

Re:not believable (1)

Yvan256 (722131) | more than 3 years ago | (#35844634)

What's worst is that the small photos on that page are 50x50 pixels, not 25x25.

Ha! (4, Funny)

Sooner Boomer (96864) | more than 3 years ago | (#35844440)

CSI can do it with only ONE pixel!

Re:Ha! (1)

Anonymous Coward | more than 3 years ago | (#35844570)

ENHANCE!

Re:Ha! (1)

Anubis IV (1279820) | more than 3 years ago | (#35844576)

Only after they hit the Enhance button.

Re:Ha! (0)

Anonymous Coward | more than 3 years ago | (#35844666)

CSI can do it with only ONE pixel!

That's about how big my dick is ;-P

Re:Ha! (2)

SheeEttin (899897) | more than 3 years ago | (#35845122)

One pixel? Hell, I could do it with one bit (assuming the bit is 0 for female, 1 for male (or vice versa)).

Re:Ha! (2, Funny)

Anonymous Coward | more than 3 years ago | (#35845468)

I can do it with ZERO bits.... with 50% accuracy!

Re:Ha! (1)

roman_mir (125474) | more than 3 years ago | (#35845596)

Not true. There is normally a reflection in that one pixel of entire scenes of crime.

Re:Ha! (1)

northernfrights (1653323) | more than 3 years ago | (#35845988)

Chuck Norris can do with with 1 pixel too. But he can also tell you what they had for lunch.

Determining sex should be easy... (1)

John Hasler (414242) | more than 3 years ago | (#35844442)

...if the right body parts are in the image.

Re:Determining sex should be easy... (1)

udippel (562132) | more than 3 years ago | (#35844966)

Not so fast, mate! There are those where some parts of the body (right?) are a bad indication of what to expect several inches further up (or down).
Make it 2x25x25 for the better.

Re:Determining sex should be easy... (1)

Anonymous Coward | more than 3 years ago | (#35845290)

May I now present transgendered people including intersexed individuals?

Re:Determining sex should be easy... (1)

Lehk228 (705449) | more than 3 years ago | (#35845528)

chatroulette already invented this algorithm

Insufficient information. (2)

Behrooz (302401) | more than 3 years ago | (#35844444)

Determine gender at what precision? TFA wasn't very enlightening... indeed, listing mis-identified faces doesn't really help much here.

This is like the problem of false positives in airport scans, but without the terrorists. :P

Re:Insufficient information. (2)

rm999 (775449) | more than 3 years ago | (#35844534)

The article has a histogram that shows how sure the algorithm was of its predictions for both sexes. Males on the left of 0 were misclassified, and vice versa for females.

Now, the only confusing this is if that plot is for the test set of the train set. If it is for the test set then it answers your question. If it is for the train set it tells us a lot less. Pretty sloppy of them to title a graph with both :(

a non-white full headscarf surrounding the face (1)

Anonymous Coward | more than 3 years ago | (#35844468)

indicates a woman with probability greater than 0.9.

1 bit should be enough (1)

rossdee (243626) | more than 3 years ago | (#35844472)

0 = female
1 = male

Re:1 bit should be enough (0)

Anonymous Coward | more than 3 years ago | (#35844700)

X=Female. Y=Male.

Re:1 bit should be enough (2)

$RANDOMLUSER (804576) | more than 3 years ago | (#35844708)

That's exactly the kind of sloppy thinking that had us "remediating" software for three years prior to Y2K. Where, in your grand scheme of things, are the values for (as examples): Michael Jackson, Lady Gaga and Richard Simmons? Please, won't somebody think of the mutants?

The Kaulitz Test (1)

amanicdroid (1822516) | more than 3 years ago | (#35844492)

How'd it categorize the lead singer of Tokio Hotel?

Re:The Kaulitz Test (1)

Yvan256 (722131) | more than 3 years ago | (#35844640)

Plastic.

Re:The Kaulitz Test (1)

udippel (562132) | more than 3 years ago | (#35844982)

ternary logic

Enhance! (1)

gmuslera (3436) | more than 3 years ago | (#35844500)

At least it works on tv

weird positioning on LDA (2)

Trepidity (597) | more than 3 years ago | (#35844512)

It's not like using linear discriminant analysis is some crazy or countercultural thing. It's a common simple technique. On some data it works well, and on such data, it's not uncommon to use it. It's particularly common in image-identification type tasks, and is one of the classic approaches to face recognition.

Re:weird positioning on LDA (1)

snarkh (118018) | more than 3 years ago | (#35844722)

That's right. Moreover, in the paper they say that for larger data set SVM with an RBF kernel performs best.

it is puzzling (2)

snarkh (118018) | more than 3 years ago | (#35844520)

that an application of a standard machine learning method can be patented. They have a publication in a good journal (PAMI), but there is nothing earth-shattering in the research. As far as the comparison with SVM is concerned, non-linear SVM does beat the linear methods when there is enough data (as they acknowledge in the paper).

Re:it is puzzling (0)

Anonymous Coward | more than 3 years ago | (#35845624)

SVM beats LDA so long as the decision boundary is not exactly at the middle between the centroids. Is there "texture" to the distributions and do you have enough samples to see it? If so then you can do better than LDA with most of the methods that allow a flexible boundary. LDA essentially relies on the same sorts of assumptions Fisher made when devising parametric statistics -- if you don't have enough samples or can't crank through non-parametric methods computationally then you make assumptions and hope nature knows to obey the rules.....

Re:it is puzzling (0)

Anonymous Coward | more than 3 years ago | (#35846168)

In my opinion, patented an algorithm that recognizes images of males (with mustache!) as females is earth-shattering.

It's Pat! (0)

Anonymous Coward | more than 3 years ago | (#35844528)

And yet, after watching SNL in High Def, most people can't figure out Pat's gender.

How convenient... (1)

fuzzyfuzzyfungus (1223518) | more than 3 years ago | (#35844530)

How handy. I think that it was just this week that the DoD was looking for research aimed at assisting them in blowing up slightly fewer noncombatants based on lousy aerial footage...

Chatroulette (0)

Anonymous Coward | more than 3 years ago | (#35844536)

I believe that chatroulette already solved this problem.

The thing about SVM.. (1)

kvvbassboy (2010962) | more than 3 years ago | (#35844538)

Quadratic programming with an RBF or Gaussian kernel should give you the best possible separation between any two classes by design, with sufficient amount of cross validation. Sadly, this doesn't always work in practice. I spent many months working on getting SVM to classify speech datasets, but the simpler methods always reigned. Not to mention, they take a fraction of the time to train a model.

I am guessing that the parameter tweaking required for SVM in some datasets is much more sensitive than others.

inb4, you don't know how to use svm

Humans can do it with 12 points of light (1)

Beryllium Sphere(tm) (193358) | more than 3 years ago | (#35844552)

I can't remember where I saw it, so can't give you a link, but there's a video of two people in a completely dark room with small light sources at joints and extremities. The instant they start moving, you can tell which one's the man and which one's the woman.

Re:Humans can do it with 12 points of light (0)

Anonymous Coward | more than 3 years ago | (#35844598)

Seems to be what you are talking about.

http://books.google.ca/books?id=lEDyN5-80E8C&lpg=PA171&ots=sTyai15toi&dq=12%20points%20of%20light%20movement&pg=PA171#v=onepage&q&f=false

Re:Humans can do it with 12 points of light (0)

Anonymous Coward | more than 3 years ago | (#35844840)

The key word here is moving. Which means >220 (eye's ability to distinguish small changes in light, according to the USAF) times those 12 points per second. In a 70 MP picture. (According to Wikipedia, that's what the human eye can do. 50 CPM. 160 deg. * 175 deg. FOV)

So we could say that we see if it's a man or a woman from
220*70,000,000 = 15.4 gigapixels.
Not very impressive. ;)

The Air Force will like this (1)

gman003 (1693318) | more than 3 years ago | (#35844564)

The US Air Force recently launched a challenge for a system "that can determine approximate age (adult, teen, child) and gender of small groups of people at a distance." [challenge.gov] , with the goal of reducing civilian casualties during UAV operations. It shouldn't be too hard to make a system that can guess ages (at least well enough for their purposes), so the research team practically netted $20k already.

Additional news coverage: http://www.wired.com/dangerroom/2011/04/boy-from-girl/ [wired.com]

Like the man said... (1)

nevillethedevil (1021497) | more than 3 years ago | (#35844582)

625 pixels should be enough for anyone

Re:Like the man said... (1)

$RANDOMLUSER (804576) | more than 3 years ago | (#35844632)

It's not the number of pixels, it's how you use them.

Re:Like the man said... (1)

Tablizer (95088) | more than 3 years ago | (#35844680)

of course, coming from people who call themselves "micro" and "soft".

SVMs vs. LDA (4, Informative)

hoytak (1148181) | more than 3 years ago | (#35844586)

The algorithm is also interesting in that it proves that an older and fundamental pattern recognition technique - linear discriminant analysis is just as good as the more trendy Support Vector Machines if used correctly and much more efficient.

A bit of clarity might be useful here. Support vector machines use linear discriminants as the central part of the algorithm. These linear discriminates -- simply hyperplanes separating two regions, are defined by a subset of the data points (called the support vectors). The other key part of an SVM is that it projects the data into a high-dimensional space in which hyperplanes can appear as curves or other shapes in the original space. This higher dimensional space is determined from the data using distances between the points in the data set (it's a kernel space).

The net result of all this is that SVMs are pretty much guaranteed to always perform better in terms of misclassification error than a simple linear discriminant, as every possible linear discriminant is considered in building the SVM. But it can be slower, and it can overfit.

So what's going on here? Linear discriminant analysis is an old statistical technique (1930s) that fixes a hyperplane based on distributional assumptions about the two classes. This allows the two classes to be plotted in a simple histogram by projecting them to the normal of this hyperplane, as shown in the picture in the article. It's used all over in statistics, and it works very well when dealing with two symmetric Gaussian distributions (that's what the theory assumes).

Thus the reason it works well here is that they've managed to transform their data in such a way that the two classes look like this sort of distribution. That's the insight here, not the choice of classifier. When the simplest model works, more complex techniques will overfit, meaning that you train on noise instead of the underlying structure of the data.

Re:SVMs vs. LDA (1)

Anonymous Coward | more than 3 years ago | (#35844638)

Well of course. Thanks captain obvious

Re:SVMs vs. LDA (1)

tgv (254536) | more than 3 years ago | (#35845758)

There are quite a few people who don't know anything about it. They don't know what an SVM is, nor what differentiates it from linear separation (aka Perceptrons). So any explanation is more than welcome, and the GP got rightly modded up. Perhaps an even more *obvious* explanation is needed. Why don't you write one?

Re:SVMs vs. LDA (1)

ace123 (758107) | more than 3 years ago | (#35844726)

Thanks for the explanation! I would mod your post up if I had points.

What It Looks Like (2)

pgn674 (995941) | more than 3 years ago | (#35844596)

Here's what 25x25 pixel faces look like, using the example from the article: Picasa Web Albums - Paul Nickerson [google.com]

Re:What It Looks Like (1)

Jane Q. Public (1010737) | more than 3 years ago | (#35844668)

Hey! I've seen that set of pictures before. They gave the machine a trick question! All those people have been neutered.

Real test would be in Bangkok (0)

Anonymous Coward | more than 3 years ago | (#35844622)

I have a friend who took someone home only to find "it" wasn't the gender he was looking for! (I, on the other hand, know to pay particular attention to uh, other parts of the body... in particular the hands).

Re:Real test would be in Bangkok (0)

Anonymous Coward | more than 3 years ago | (#35844748)

meh, an ass is an ass and a mouth is a mouth. It's not gay unless you bottom.

Even better (0)

Anonymous Coward | more than 3 years ago | (#35844704)

I can do it with 0 pixels and 50% accuracy.

But.. (1)

XB-70 (812342) | more than 3 years ago | (#35844734)

I think he means gender. Identifying sex is much, much harder. Are two people who are hugging having sex?

Re:But.. (1)

yuri benjamin (222127) | more than 3 years ago | (#35845258)

"Gender" is a grammatical term. "Sex" is a biological term. So yes, he did mean sex.

Are 625 Pixels Enough To Identify Sex? (1)

Odinlake (1057938) | more than 3 years ago | (#35844772)

And here I thought the title must relate to some kind of automated video analysis, you know, what-is-porn-what-is-not.

I believe ... (1)

PPH (736903) | more than 3 years ago | (#35844948)

... Mick Dundee's method is more reliable.

Low bandwidth porn (1)

erice (13380) | more than 3 years ago | (#35844956)

I would have thought you would need more than 625 pixels. Must have been some interesting research.

What? What do you mean: RTFA? I get all the information I need from the titles!

Saturday Night Live: (1)

Hartree (191324) | more than 3 years ago | (#35845116)

Maybe we'll finally answer the question in the skit theme song.

"Is it a man, or is it a woman? It's Pat!"

Bet it can't tell a dog from a cat (1)

cullenfluffyjennings (138377) | more than 3 years ago | (#35845142)

with any number of pixels. Color me a bit skeptical about this ... often when one looks at the training data used to train it and the test data used to test it, much is revealed about how it works.

not AI at all... (1)

metageek (466836) | more than 3 years ago | (#35845514)

calling discriminant analysis an"old AI method" is like calling a typewriter "an old terminal".

Discriminant analysis was invented by Fisher and it is clearly a statistical method. The term AI would take another 20 years to be coined...

Fuck no (0)

Anonymous Coward | more than 3 years ago | (#35845666)

Fuck no they aren't enough pixes to detect this [sankakucomplex.com]

Aerosmith provides some commentary (0)

Anonymous Coward | more than 3 years ago | (#35845898)

Aerosmith provides some commentary on the limitations of such a system...
http://www.youtube.com/watch?v=nf0oXY4nDxE

Statistical anomalies (2)

Gaygirlie (1657131) | more than 3 years ago | (#35846024)

These things can never become truly 100% perfect as there's lots of people that will show up as statistical anomalies. There are for example people who suffer from hormonal imbalancies resulting in overly feminine looks in a male, or overly masculine looks in a female. Just as well transsexual people will be hard for these things: hormonal medication does not change skeletal features, but they change distribution of fat in the body, including face, and thus for a machine they'll like fall in the grey area between either gender. And how about intersexual people who are physically neither gender? I had a friend before who was IS and it just was really hard to tell from the looks what gender one should assume. Mentally she identified as female, but that can't obviously be told from a picture.

This also makes me wonder about the future.. I hope these "gender guessing machinery" do not become the norm in our society and public areas because they will lead to lots of issues with the aforementioned groups of people.

Re:Statistical anomalies (1)

BlueParrot (965239) | more than 3 years ago | (#35846400)

Just as well transsexual people will be hard for these things: hormonal medication does not change skeletal features,

Even so quite a few transsexuals undergo various forms of plastic surgery that certainly can change a lot of skeletal features. This can range from rhinoplasty, forehead contouring, chin reductions etc...

Re:Statistical anomalies (1)

Gaygirlie (1657131) | more than 3 years ago | (#35846434)

Just as well transsexual people will be hard for these things: hormonal medication does not change skeletal features,

Even so quite a few transsexuals undergo various forms of plastic surgery that certainly can change a lot of skeletal features. This can range from rhinoplasty, forehead contouring, chin reductions etc...

Only rich ones do, those cost helluva lot of money. I have one FtM and one MtF friend, neither of whom can afford such and are only on hormones, and belonging to a sexual minority group myself I hang out a lot in one of the local forums for HLGBTI people and so far I have not met anyone else either who would have had anything else done than hormone therapy.

Obligatory joke... (1)

Altesse (698587) | more than 3 years ago | (#35846198)

... 625 pixels should be enough for everyone !

Xkcd ruined my life (2)

alendit (1454311) | more than 3 years ago | (#35846256)

Had think about this somehow http://xkcd.com/598/ [xkcd.com] .

ZOOM and ENHANCE! (0)

Anonymous Coward | more than 3 years ago | (#35846282)

Zooming zooming zooming.. Enhancing.. There we go. Its a .... a .... male pixel.

25x25 pixels may work for ID-ing Spanish men (0)

Anonymous Coward | more than 3 years ago | (#35846452)

...but here in America we need at least 28 pixels LOL

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>