Image search is a hard problem. That is especially true when you are searching with no information other than the image itself (no tags, titles, or descriptions, just the photo). I’ve seen my fair share of image search demos, and they usually promise far more than they deliver. But last week, I finally saw one that deserves the name. It is called TinEye. The first 500 readers to send an email to techcrunch [at] tineye [dot] com will be pre-approved for the private beta.
You can upload any image file to TinEye or paste in the URL of an image in the search box. It then will scour the Web and find sites that contain the same image (see video). As long as the image has been indexed by TinEye, it can be found. For instance, it identified 39 sites that use the famous Steve Jobs image at left and in the screen shot below. It only launched earlier this month, but so far TinEye has indexed close to 750 million images on the Web, and should reach one billion sometime next month. Google, in comparison, is estimated to have indexed between 4 to 6 billion images. (Polar Rose, on the other hand, has only identified 21.5 million images on the Web, although that is more facial recognition than exact image identification).
TinEye works even if an image has been modified, cropped, made into black and white, or photoshopped. The image-search algorithms behind it were developed by a company in Toronto called Idée, which already offers a service to news photo agencies such as the Associated Press, Agence France-Presse, and Getty Images to automatically identify whenever one of their photos is used in a newspaper or magazine. Idée digitally scans thousands of print publications for this purpose (before, it was actually someone’s job to leaf through every page looking for the photos). Digg also uses the technology to make sure no duplicate images are submitted to its image section, and Adobe licenses the technology in Photoshop for the “find by Visual Similarity” feature.
At this point TinEye still is not much more than a technology demonstration. After all, most people search for images by putting in a keyword that describes it, not by showing it the image they want. (If they already have it or know where it is, why do they need to look for it?). Its main appeal right now is for photographers and photo agencies who want to find out where their photos are being used on the Web. (Think Attributor).
But you can imagine other ways the technology could be applied. If Idée can figure out a way to seed the images it finds with tags, or combine its approach with a text-based index, it could create an image search engine that is really good at finding exactly what you are looking for. Unfortunately, it is not good at finding images that are similar to each other (see update below), and that could limit its potential. Personally, I’d love to see Apple license something like this and incorporate it into the next version of iPhoto. Add an auto-tagging function, and it would be able to find all the pictures of, say, my sons and automatically label them with their names. I’d actually pay extra for that feature.
Update: I was mistaken about that last bit. TinEye canot do facial recognition or find similar images, it can only find the exact same image. So it wouldn’t work as a way to auto-tag pictures in iPhoto. I believed that it could due to a miscommunication with Idée’s CEO (my fault). The original title of this post was “Mr. Jobs, Here’s An Idée For You: Put TinEye Image Search Into iPhoto (500 Invites).” I got a little carried away there, but I still want that capability in iPhoto. TinEye is just not the answer.
This can also be useful for searching for higher resolution photos of an image you already have.
Looks Interesting….i was thinking of an idea like this about a week ago..ironic…sigh there goes my chance of a start up
I had the treat of a full tour of Idée’s headquarters in Toronto last week, amazing space, amazing talent and amazing technologies…
The Beta feels complete and is amazingly effective and accurate, even finding actual photos of my cat, of which I have no idea how they ever were released to the web.
Worth trying !!
Thats cool..
Here’s an idea for the iPhone: put a built-in webcam on it and load it with skype. A total jetson’s video phone!!!
Heck, put 2 built-in webcams — one pointing toward the face of the caller, and one on the opposite side (the back of the phone) pointing at what the caller is looking at. that way they can be talking with a friend and then say “check this out!” and hit a ’switch views’ button that changes which camera it transmits to the receiving party….so they can see what you’re looking at. then switch back to look at your face again.
Do that, with huge memory, good battery life, seamless integration with Exchange, and a 7 megapixel camera with movie mode, and I think it’s about the ideal device.
of course, it’s easy to sit around & tell others what to build when you don’t have to do it yourself, haha!
This looks pretty sweet. Looking forward to an invite… I may even use it in twitturly for an idea I had.
cool idea.. not sure how I would use it. btw, why do I have to be invited to do a beta search?
thanks, i will wait u come out of beta.
It also has some good uses in education. Because it will also find variant/derivative images as well it allows for some interesting lessons. I blogged about this a couple of days ago http://nwinton....8/05/24/tineye/
But this is picture recognition, not facial recognition so it has nothing to do with what you suggest — Bonne idée!
that’s great technology – can’t wait to see how it works. i’m sure google could put this to great use as well, as well as any other broad search platform
Its great to see Leila and her team start reaping the fruits…I met her last week at the mesh conference in toronto and they fully deserve the coverage with such a solid product. Can’t wait to see more great things for them in the future.
I wouldn’t actually call this “image search”, anymore than uploading the entire contents of an article into a Google Search box, and asking it to find pages with most of the same text is considered search.
Moreover, it’s a much much more tractable problem to solve than the general image search problem, which entails feature recognition. I don’t know which technique they’re using, but finding image matches in the presense of linear transformations and noise is a mostly solved problem, so I’m wondering if they really invented the algorithms themselves, or just lifted them from a journal paper.
That’s a very Apple-like demo; Like they’re trying to appeal to them by saying “Look how consistent our brands are!”
Perhaps they already have their eyes on that licensing deal
re:#1
this is the only use i would use this for. its great, but yes, why would i want to look for an image i already have? heh
@Tim, I thought the same thing, but the CEO told me that it could be used for a desktop photo management app as well.
I like Mr.Jobs,He wages is 1$
Now this is cool
Find out if your photos are being used on fake myspace/facebook profiles and go after the culprits!
TinyEye should have a big impact on the web, just think, now people will have to be worried about stealing pictures. TinyEye might drive greater addoption of CreativeCommons photos or show people that it is impossible to control an image once it is online… either way… I can’t wait to see the effects…
I believe you are Talking about TinEYE, not TinyEYE!
@Erick, like Powerset can achieve natural language search on the entire web? Sure. I bet Steve Jobs is writing a check right now.
To be clear, you’ve seen attempted but flawed facial and feature recognition systems partially work, but you weren’t impressed. You didn’t see this do anything you particularly need, but you believed them when they said it could do anything? Great, thanks.
@Tim, not anything. Something similar to what I saw, which was a more in-depth demo beyond just TinEye.
But I have another e-mail into the CEO to double check that the scenario I outlined would actually work. We’ll find out in the AM.
Leila the CEO is a great Canadian entrepreneur. Active in the startup community and runs a great team.
I had the pleasure of touring their Toronto office a couple of weeks ago and it is an impressive operation.
The company is earning good revenue and without going into confidential info, they are in a great position in a very interesting market.
This is just a technology demonstration, but here are major industries that depend on image recognition and search and many more that could build profitable businesses of working image search.
It will be exciting to see where the company takes this from here.
Cute girl on the video.
Sounds like a CIA Spooks technology!
I am sorry Sir, you cannot enter USA because a picture of you is found as an Avatar of Anti-Americanism forum, Playboy forum, and Pravda newspaper!
But Madam, does that make me undesirable? Sir we are only doing our job and protecting Freedom and Liberty of the American people!
The Big Brother is watching You!
Or maybe a little more rudimentary! Hey I like this guy, but let me do an image search on his face and see who were his ex girlfriends!
Tim, you are right that picture recognition is not facial recognition.
We (Luxand.com) work out facial and feature recognition systems and its much harder task.
@Erick. No, it won’t work that way. This doesn’t search the content of your photos. If you have a picture of a car you saw, it won’t tell you what car it is, nor will it tell you that you have a picture of a car… unless that exact picture was used on website. So another words, Mr. Jobs, ignore this post!
Their technology is the aspirin for some photo publishers’ headaches.. but i don’t see it useful for average folks and it won’t do what google does for search, not yet anyways.
Maybe this is good for social networks to reduce photo redundancy…
COOL. That’ll show my art history teacher…always making me identify photos that aren’t even in the book.
Will it find the same image if the image name has been changed? Sounds to me like that’s actually what it’s doing…
Erick, I don’t think this would help you find a group of pics of your son. From the caps, it looks like it will help you find different versions of the same picture.
I can see why news agencies would love this technology, but don’t think home users would see much purpose in it.
WOW!
Cool technology. Is this another company that will struggle to find a business model just like Riya did ?
Would like an invite for TinEye, if you please.
@32
…send an email to: techcrunch [at] tineye [dot] com
The lovely Amber MacArthur
I tried TinEye recently. It finds copies of images and does a good job. It finds even ones that are modified – noise, color, stretch, crop, some photoshopping. It doesn’t work as well with rotation. That’s a major drawback (compare to Lincoln from MS Research).
@Tim and others, you were right about the facial recognition part not being part of the technology. Thanks for pointing that out, and my apologies.
The CEO confirmed what you said this morning. I’ve updated the post and changed the title so that it is not misleading.
I’ve got to be more careful. about those late night posts. But at least I’ve got readers like you guys to keep me honest.
I hate stuff like this where it is restricted to an invite only.
What’s the point of showing me what I cannot actually use!!!
Flip the image horizontally and the algorithm goes “boink”
This could be a good tool for a brand / commerce website to see who has swiped their imagery and is using it to re-sell their product. It’d be nifty to test that out — TinEye — shoot over an invite!
his is a great demoware!
But if you want real visual similairty engine + award winning facial recognition goto http://www.picitup.com …
iPhoto could get content-based image querying features if open-source code from http://www.imgseek.net/ is plugged into it.
I’ve been lucky enough to use the beta for a while now. Like ozz314 says, it’s a great way to find higher res images. It’s also a great way for our startup (GigPark) to track people talking about us…. I just upload our logo and I can find where it is used on the web. We use it like a blog search for images.
Very cool technology, great people and huge potential – I really like Idee!
excelente
so does anyone know how to get an invite? i need one bad! this sounds like an awesome program but i havent even tried it out and im DYING to
Hi Ewan,
You can get an instant invitation on the idee blog – http://blog.ide...na-lisa-widget/
There are 500 up for grabs. Check at the bottom of the post!
Cheers!