109
$\begingroup$

I'm working on a little project involving the faces of twitter users via their profile pictures.

A problem I've encountered is that after I filter out all but the images that are clear portrait photos, a small but significant percentage of twitter users use a picture of Justin Bieber as their profile picture.

In order to filter them out, how can I tell programmatically whether a picture is that of Justin Bieber?

Improve this question
$\endgroup$
9
  • 19
    $\begingroup$ What's your development platform? This can easily be done in .NET because it superior over all other programming environments. Simply call the Page.EradicateBieber() function. Microsoft foresaw this need and graciously provided it for us out-of-the-box in .NET 4.5. (Those of you on older versions will have to wait.) (That is, of course, all tongue-in-cheek.) $\endgroup$
    – David Stratton
    Feb 14, 2011 at 22:46
  • 35
    $\begingroup$ I think I can safely assert that SO doesn't need a [justin-bieber] tag. $\endgroup$
    – skaffman
    Feb 14, 2011 at 22:47
  • 2
    $\begingroup$ I can safely assert that people spend more upvotes on comments and this question than the close option (the one that deserves the votes). $\endgroup$
    – Linus Kleen
    Feb 14, 2011 at 22:50
  • 20
    $\begingroup$ A Justin Bieber audio filter would be good too $\endgroup$
    – Paul R
    Feb 14, 2011 at 22:58
  • 6
    $\begingroup$ stackoverflow.com/questions/953714/face-recognition-library $\endgroup$
    – Ami
    Feb 15, 2011 at 0:54

7 Answers 7

Reset to default
49
$\begingroup$

A better idea might be to trash all images that appear in the feed of more than one user - no recognition needed.

Improve this answer
$\endgroup$
5
  • 3
    $\begingroup$ Yeah, maybe set a threshold of 2-4 possible duplications (to handle the new-baby case) before you reject a photo. Depends on what you're going to do with the photos, I guess. $\endgroup$
    – Mark Bessey
    Feb 15, 2011 at 1:49
  • 3
    $\begingroup$ Simple, elegant solution. +1. $\endgroup$
    – Robert Harvey
    Feb 15, 2011 at 3:40
  • 14
    $\begingroup$ People could use different pictures of the same person. $\endgroup$
    – Rebecca Chernoff
    Feb 15, 2011 at 3:47
  • $\begingroup$ (+1) at Rebecca and (-1) @ PPPPPP: This just shifts the problem. $\endgroup$
    – steffen
    Feb 15, 2011 at 7:15
  • 5
    $\begingroup$ They could, but in most cases they're gonna choose from a relatively small pool of images, so it'd probably still work. Edge cases be damned - for all you know my picture is of my uncle anyway. $\endgroup$
    – naught101
    Mar 21, 2012 at 7:05
16
$\begingroup$

I have a feeling that this may be the solution here.

Simply throw the Twitter profile image to Tineye, see if it returns images (and associated URLs) that can clearly be identified (or automatically scored using simple word-count logic) as being related to (or of) that little sack of ****.

Simples!

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Google announced the image search recently: youtube.com/watch?v=t99BfDnBZcI I don't know if it has an API yet, but that may be an alternative. $\endgroup$
    – petrichor
    Jun 16, 2011 at 10:30
12
$\begingroup$

Since you are able to filter to only those that are clear portrait photos, I'm assuming you have some method of feature generation to transform the raw images into features that are useful for machine learning purposes. If that's true, you could try to train a classification algorithm (there are lots of them: neural networks, etc.) by feeding the algorithm a bunch of known Bieber photos as well as a bunch of known non-Biebers. Once you have trained the model, it could be used to predict whether a new image is Bieber or not.

This sort of supervised learning technique does require you to have data where you know the correct answer (Bieber or not), but those could probably be found from a Google image search. It also requires that you have the right sorts of features, and I don't know enough about image processing or your algorithm to know if that is a major drawback.

Improve this answer
$\endgroup$
4
  • 2
    $\begingroup$ Unfortunately, the feature generation step is both the hardest and most important one :(. $\endgroup$
    – steffen
    Feb 15, 2011 at 7:19
  • $\begingroup$ @steffen There is some suggestion that the OP is messing with the faces, so have some descriptor generator. $\endgroup$
    – user88
    Feb 15, 2011 at 12:15
  • $\begingroup$ @mpq: I did not doubt that, however, if the OP does not have one feature per pixel, then he has to find a meaningful aggregation level. I did not downvote, I just wanted to point to the complexity which lies behind this answer (which is, of course, correct). $\endgroup$
    – steffen
    Feb 15, 2011 at 13:00
  • 1
    $\begingroup$ Right, the feature generation step is the hard part. I was assuming OP could do this since he has some mechanism for processing the images already. Even if he does though they might only be useful features for detecting face/not face instead of Bieber/not Bieber...it really depends on the features. $\endgroup$
    – Michael McGowan
    Feb 15, 2011 at 13:17
7
$\begingroup$

You could use a method like eigenfaces.

The following link has a good walk through of the procedure as well as links to different implementations. Link

From here it is common to use this in a classification approach, train a model and then predict cases. You could do this by training on a bunch of known celebrities and if you predict a face from twitter as one in your trained model of celebrities, remove it. Similar to this.

This suffers from constant amendments. Soon there will be a new Justin Bieber that wont be in your trained model, so you cant predict it. There is also a case like Whitney Houston, you may have never thought to add her before but she may be a common image out of respect and admiration for a few weeks. You will not have the downside of baby pictures as mentioned above though. To over come these problems you could use more of a hierarchical clustering approach. Removing the first few sets of clusters that are very close if they reach a certain level of support, your first cluster has 15 items before a second is constructed. Now you don't have to worry about whose in your training model but you will fall to the baby pictures issue.

Improve this answer
$\endgroup$
5
$\begingroup$

If you want to do it yourself, I would recommend using Intel's free and open source OpenCV (CV for computer vision) project.

Link.

Link.

Improve this answer
$\endgroup$
3
$\begingroup$

You need to put on an algorithm detecting which person that picture is referring to. You can build a model based on different portrait pictures of famous personality and use classifiers to ensure that this picture is referring to one of your database picture. You need to use a certain classifier based on different parameters liked to the face, like distance between eyes or other parameters to rise up the accuracy of your model. There is also skin analysis. The most important is to build a good classifier. This method can be vulnerable.

But there is also a very good project working on face recognition: link.

Improve this answer
$\endgroup$
1
  • 3
    $\begingroup$ AFAIK OpenCV and the linked site only implement face detection (ehere in the picture is a human face?) which is only a first step towards face recognition (whose face is it?) $\endgroup$
    – f3lix
    Nov 20, 2011 at 15:48
2
$\begingroup$

You could try locality sensitive hashing.

Improve this answer
$\endgroup$
1
  • 2
    $\begingroup$ Plain KNN is not very good for faces. Faces have been shown to lie on a ~25 dimensional nonlinear manifold of the pictures. $\endgroup$
    – bayerj
    May 11, 2011 at 19:05

Not the answer you're looking for? Browse other questions tagged or ask your own question.