We’re getting information about a new stealth startup called Vutool. The founder is Sebastian Thrun, Director of the Stanford AI Lab. We hear that he has a team of ten or so people, most of them Stanford Ph.D. students.
Vutool is a “Google Earth” from the ground level. It is being created the hard way - a battalion of cars are traveling around major U.S. cities and taking pictures at ground level, with 6 degrees of rotation. The company calls the experience “street level immersive imagery.” Images are captured from every possible angle - the default view is as if you are in the driver’s seat.
In principle it sounds a bit like Microsoft’s Street-Side service, but on a much grander scale. Where Street Side has only a few photos every block or so, Vutool is a more realistic virtual world. One source described it as “like grand theft auto, but the real world.” Huge amounts of intellectual property were reportedly developed for the project, particularly around image processing, filtering, storage and rendering.
But the service may never launch. They were talking to a number of Silicon Valley venture capitalists, looking to raise $18 million or more in a Series A round of financing. But after receiving a number of term sheets Thrun backed out, saying the company had been acquired.
There are strong indications that Google is the acquiror, and will eventually pair Vutool with Google Earth. Users could zoom from street level to satelite view at will.
Google Metaverse may be here sooner than we think.
The vutool.com domain currently redirects to http://dent.stanford.edu/flash/start.html. Vutool.com is registered to Thrun, but there is little other interesting Whois information. Thrun did not respond to an email requesting comment.
I think on open, user contributed, Wikipedia-style approach for this stands the best chance of succeeding wildly.
I guess that would set a new record: get acquired *before* you launch:-)
Google’s taking a lesson from Yahoo’s screw up when it passed GOOG for few mil.
10 Stanford PhDs…must have scared Google, at least for a few milliseconds.
-Zaid
Make that few bil, not mil:)
Excite passed on GOOG for $1M
vutool sounds great, what Gabe said about user-submission is interesting too. Eventually most cameras (be they on a PDA, phone or just ordinary) will have GPS and will be able to embed co-ords into an image. They will just need to scoop all these images up and put it all together from there (not sure how they would work out which direction you are facing though)
Sounds like Aspen Moviemap at MIT (from like 1978)
http://www.naimark.net/projects/aspen.html
Wasn’t Amazon doing something similar as well? I guess it’ll be cool when whoever does it, but somebody’s got to actually finish sometime.
The beauty of street-level pics is being able to see buildings, signs, business, etc. For anything else, closeup satellite view is good enough.
A few years ago I did that same thing in my hometown. Today half of the pictures are clearly outdated (bussiness moved, new construction, etc).
Now do that for all major US cities and when you’re done taking pictures, you have to start again for at least 1/8 of them.
For user submitted pictures we already have Panoramio. Agree it’s not the same thing as Vutool, it’s actually better IMHO.
I think this will take some time to cover other countries or maybe they will just restrict this to the US.
Amazon A9 has a cool feature that shows the storefronts in major cities. You can walk up and down the block.
I agree that our collective photos and videos will be fascinating when we have GPS coordinates assigned to each (in addition to the usual date and time).
Is it possible that Yahoo is also moving in this direction? Geotagging photos on Flickr…Flickr has got photos of pretty much every place you can think of, but its just about how to pinpoint them down a few more levels to buildings.
thanks to anthropocentric for reminding me it was Amazon A9… i was in the same boat as Kewtr — knew Amazon had something, but couldn’t remember what it was.
looking forward to the metaverse!
Thrun led the Stanford team that designed Stanley, the autonomous vehicle which won the Darpa Grand Challenge a couple of years ago. The above mentioned image processing and filtering from a vehicle mounted camera to render a realistic world sounds like it could be an offshoot of some of the work on Stanley and the current autonomous vehicle they are designing.
Very interesting project indeed!
Take a look at http://relive.atlasct.com/ from AtlasCT, they did something nice with location and images content.
@Paul - mobile crunch also has a post on relive http://mobilecrunch.com/2007/0.....le-photos/
“Stealth startup”? Aren’t most startups “stealth” by definition until they get publiciity?! “pssst, Hey mister, Wanna buy some air?”
You could say that Chris. Usually though they launch before they get bought. Or at least put up a landing page.
Not really, at least in the netherlands we have company which is already succesful with these kind of services,
http://www.cyclomedia.nl/index.php?lang=en
With regards to outdated pictures… How about making he service “nearly-live”? For example, in London, there’s zillions of CCTV cameras, I believe doubling the number would pretty much cover everything. Don’t think same is in other cities, but you get the drift…
Virtual Tour operators would be queuing to tap the opportunity.
http://www.tekno-world.blogspot.com
Good timing on this article. It explains the car I saw driving down Cesar Chavez street in San Francisco yesterday. It had a camara with four lenses on top of a very long pool mounted on the roof. Made me exclaim, “What the hell is that?”
I also recall reading a couple years ago that Amazon had people driving around in SUV’s with cameras and hard-drives stacked in the back seat indexing images.
I agree with the poster who said the best way to make a true visual map at street level would be wiki style, where folks can upload their own photos.
an amazing - idea - if ever completed.
- I wonder if their is a business model besides “sell to bigger company” ..
- Rb
–Then again if it works…
There’s this as well, based in South Florida: http://360inaction.com/
Interesting name; don’t think it’s a great name. Wonder what’s the idea behind the name, if any.
Back in the 90’s in France, the French Post Office created exactly what Vutool is about to put together here, and it is used by 50 million Frogs like me. You can check it for yourself at http://ved.pagesjaunes.fr/ciwe.....ilPhoto.do
Odd enough, I am about to give you guys the possibility to build all the shops, restaurants, offices, etc you have in your street with an open source set of tools and the sources for tens of 3D spaces like shops, offices, restauraunts, sports venues, theaters, including a complete Las Vegas in 3D. It’s gonna be the 3D-pedia we all dreamt about. Geeks will be able to make money building 3D stuff for others, while Dummies will simply pick the 3D model that best represents the shape of their room (apartment, shop, office, restaurant, etc) and upload the pictures they took from their own walls.
As I was teaching that 3D technology at the UCLA Virtual Lab with Professor Bernie Frisher back in 2001 I nurtured the goal of putting together the French Post Office concept with my 3D realizations, so that you can not only zoom from space with Google Earth (or Microsoft Earth) to ground level and immerse yourself at the street level with a system like the French PO or Vutool, but you can click on the shop or building you see at the street level and enter the virtual building or shop in 3D, recreated with the wall textures/pix people would upload along with over 2,000 free 3D objects in stock (this is no Second Life).
The great thing with our techno is that the client is just 1 MB (compare with SL, Kaneva, Blaxxum, etc…), has all the dev tools ready and tens of 3D models to start from to build the grid in true open source.
There is everything one can dream about with a 3D environment: webcam and audio chat directly in 3D (see you, SL…), yes the avatars speak live “a la” Skype and you can put the live webcam stream onto your avatar’s head. Live video can be projected on any surface or 3D object, objects and textures can be added or edited on the fly directly and live from your 3D world, etc, etc. And it works great even for people with just a modem, which truly opens the way to a real Wikipedia-like 3D project where developing countries can participate even if their installed base of PC’s is quite old (PIII, etc).
The Second Life is great for having evangelized the Internet 3D in the last 12 months (since they’ve got plenty of cash) but it just copycats what the French studio Canal Plus did in 1997 with this technology (Le Deuxieme Monde, a complete Paris in 3D where people could just do what SL did five years after).
Yes, we are in stealth mode, beta etc, although this post will probably lift the veil a little on what’s “cooking” here.
Hi Michael,
I am the CEO of city8.com. We have lanuched silimar 3d interactive map service with real local view in Shanghai, and Beijing since July, 2007.
Please visit
http://en.city8.com/panosearch....._95E8.html to have a try on the english verison.
Bruno,
I thought Visiocity was built by the French Yellow Pages, not La Poste. Am I wrong?
There is another rumored to be another well backed startup with the exact same concept in based in NYC. Might already has NYC / Boston / DC / SF already mapped out with special HD Cameras / down res’d for consumer version. Video captured via special rigs mounted to a comercial van. Sweet consumer interface for web-browsing and a more professional version that is run via a desktop app. Totally linked up with GPS with a mashup to maps.google api. vutoo is too late.
The sloshing sound of back paddling is more amusing than the memo itself. Can they possibly be so bold and, on the flip side, so naive as to spin this fiasco as an exercise in “Radical transparency”
Nice work on asking them for your file!!! Check mate! Keep us posted if you have any response!!!
Or maybe they’d be sitting in their offices right now ‘creating’ seemingly unsanitized dossiers for Om Malik, Robert Scoble, etc. I wouldn’t be surprised if they threw in a bunch of interns to go about fulfilling a fake act in ‘Radical transparency’
@5 Nik - The direction/orientation data could be derived from shadow angles, glare, and lens flare since the timestamp is also embedded in an image. You probably can safely assume that the camera’s clock will be set via the GPS.
Thank you, Sebastien for the opportunity to clarify this (see above #29). France Telecom’s has always been the owner of the Yellow Pages in France (Les Pages Jaunes). However, it is true that France’s ‘La Poste’ doesn’t share revenues with France Télécom anymore, so it has its own directory for finding people, busineses and things with street addresses in France (L’@nnuaire with @ standing for “A”…)
So, France Telecom uses MappyPro Visiocity, a software that precisely creates a database of all streets and buildings of a city with photo archives per address, with zoom and pano “a la” Quicktime VR. The company that created that software is Mappy SA, 47 rue de Charonne (75011 Paris - France). If anyone is interested (Allo, Google?…) you can contact Mrs. Virginie Tauzin in France at +1 4807 5851. I have nothing to do with them, but I hope this helps.
By the way, today starts the “Virtual Worlds 2007″ conference in NY, the first major one to specifically promote marketing in virtual worlds to Fortune 500 companies (http://www.virtualworlds2007.com). I wish I had time to attend, but there’s too much for me in LA right now
Good idea Amazon! errr Standford!
I do recommend you check the demos offered by http:\\www.360inaction.com - got my preference from all I could check.
How many PhDs does it take to take a picture? (There must be a joke in there somewhere.)
@Bruno,
Your start-up sounds pretty cool GL, shoot me an email (link in name) when you need beta testers.
I run an early stage startup called EyeStride (www.eyestride.com) which seeks to do something similar to VuTool, but we take a user-generated content approach to creating street level environments that Gabe and several others mentioned. We’re also tackling this problem from the perspective of creating an advertising vehicle that any business with a physical space can use to link that presence to a map.
Using EyeStride anyone with a digital camera can create a “tour” that anchors images to a Google Map or any user contributed map (any image can be used as a map and layered among Google maps). For example we’ve created tours of homes that start on a Google Map and then move through different levels of a house using floorplans as a navigational guide.
Sometime next week we’re going to launch the first samples of our street level tour experience and in a month or two we’re launching a limited alpha of the full “create-upload-explore” solution. You can sign up on the site
Check out our site and feel free to contact me michael {at} eyestride.com
=)
Take a look at Immersive Media. Since 2005, 360° georeferenced spherical video has been captured of North America’s major cities.
Those interested in street level imagery may find NZ company Roadworks Online of interest.
Roadworks help mainstreet and town centre groups find their place on the web. Most sites, such as Ponsonby Road feature Streetscroll. This allows users to take a virtual walk along the street, with business info displayed on mouse-over. Interestingly, these sites have been around since 2000!
The company is currently working on version 3 of its community model, due for release later in 2007. Among other modifications, this will incorporate geocoding of locations, linked to business listings, etc.
The imagery is pretty good. Certainly better than A9’s effort and presumably created for less than the billions mentioned above!