Chinese Photo Site Tuyuan May Do Facial Recognition. Or It May Just Put Boxes Around People’s Heads
by Michael Arrington on April 16, 2008

I can’t really say much about Tuyuan yet since it’s in Chinese and there isn’t much information (translated page here), but it certainly looks like they’re trying to tackle the facial recognition problem that has destroyed many a startup.

We’ve seen Riya (now focused on ecommerce via Like.com), Ookles (never launched), and Polar Rose (in private beta for nearly a year), among others. Most recently Tagcow came on the scene, but it turns out it uses humans to tag photos, which tends to produce bad data.

Will Tuyuan be any different? We have no idea yet. But we’re contacting them to find out. More soon. Thanks for the tip, Orli.

Comments

Tuyuan’s AboutUs page translated by Google –
http://www.google.com/translat.....angpair=zh|en&hl=en&ie=UTF8

Interesting service… :-o

 

According to what I can read from the website (I can read some Chinese), the website claims to automatically recognize people in the photos, and after “naming” the people, any new photos with the same person will automatically be tagged.

Sounds like riya.

 

Such a thing wouldn’t be hard for most photo’s on myspace, just put a box around the middle of the photo and you have all the heads ;)

 

if anyone can get into the beta, please provide a full report. :-)

 

According to google translation of their about-us page, it’s a “geopolitical map” LOL… :-)

 

Don’t forget Viewdle, who recently won the Plugg award for best start-up, they specialize in facial recognition, although they admittedly focus on video rather than images.

Their technology is very interesting, if only for the fact it is built around an old technology used by the KGB.

http://www.zoomz.nl/archief/pl.....y-viewdle/

 

Face detection can be performed very well with modern computer vision algorithms. Face recognition is a much harder problem but good algorithms that work for small data sets are available. From Kai’s comment it sounds like they are performing recognition within a small set of people so it is possible that they have a good solution. What I am not so sure of is whether this is really useful to anyone!

 

was my comment just deleted?? Don’t take it personally, Mike, I just think you could have waited for their response before posting this article (which basically has no content other than “I wonder what this does”).

 

According to tuyuan, it just take a few uploaded photos of the same person for the algorithm to work… seems fun..

 

This is their blog, more information. http://blog.tuyuan.com

 

I’m a native Chinese and one of Tuyuan.com’s beta testers. Tuyuan does use face detection technology to help tag people in photo and performs awesome!
For English speeking people, u can check out an example based on hundreds pics from the TV-show Friends. (photos of Monica detected automatically by Tuyuan: http://www.tuyuan.com/photolis.....mp;uid=970)

U need invitation to join tuyuan. Just submit ur email address and get surprised. LOL

 

Facebook should also add such facial recognition boxes. :)

 

startups != science.

Anyone with a university degree in cs, math or physics and appropriate coding skills can solve that problem.

As soon as startups are based on competence, and not greed they might even succeed at that.

 

There is Olaworks in Korea. They had similar issues that most in this area have, that is “where do we apply the technology?”. When I talked to the CEO in this February, they were working with ecommerce and news sites. Gradually they seem to be figuring it out.

http://www.olaworks.com/

 

The “About” page is bloody funny !!! lol

 

Does’t http://mugr.com/ do all this kind of stuff with face recognition? Any infos on that?

 

http://www.seekpai.com
there is another website providing similar function.

 

My analysis tell me they are using human more than automated technology if they ever exist in first place.

Why ? Just take a look at the box that highlight the head ?
Anyone notice that it is inconsistent ? Why some box touch the head, and some leave a noticeable distance given that the background is homogeneous ? Only a human could produce such inconsistency. A automated system tend to be more consistent and accurate.

Well, just my opinion, I could be wrong but then human judgement is inconsistent, isn’t it ? <:^]

 

Michael, you are deceived by the page shot.
i’m pretty sure this site (TuYuan) has nothing to do about face recognition.
It’s kind of a mixture of flickr and facebook sns.
btw: I’m native chinese.

 

From their official blog:

Tuyuan help in English. Part 1: From Sign Up to Upload
http://blog.tuyuan.com/archives/25

Tuyuan help in English. part 2: Facial recognition
http://blog.tuyuan.com/archives/26

btw, mugr appears to do something quite different, and seekpai stopped their effort on facial recognition a while ago.

 

looks like they have a blog http://blog.tuyuan.com/. They have some translation there.

 

With population China has it could be even cheaper to use humans than to use Amazon Mechanical Turk :)

 

Hi, Michael,

Seems like you haven’t got my emails. I sent you two emails and one invitation. If none of them reaches you, please let me know and I will figure something out.

To make some clarification, we find faces on photos, we also try to figure out who is who (facial recognition), and we don’t use human power.

For others who are interested in the service, please read our help series (3 posts by now) in English at http://blog.tuyuan.com. If you think you can still handle the language, please apply on Tuyuan homepage and we will send you the invitation soon.

 

According to the homepage screen shot, it looks like they only differentiate between foreigners & asians. =)

 

Automatically recognize people in the photos is possible. If you need the engine for your project just contact us http://www.Luxand.com

 

woo~I’m a user of tuyuan.

flying by,flying by~

 

I think I found the hidden link to their registration page, hopefully they will open the registration to the public soon. Try it yourself.

http://deploy.tuyuan.com/reg?a=reg

 

Leave a Reply

Create a Gravatar for your comments.
« Back to text comment