Presentation on our 3-month research and prototyping project in augmented reality for mobile phones. Presented at MEIC5 event (Mobile Experience Innovation Center) at the Ontario College of Art and Design, Toronto, Canada. November 26, 2009.
7. How does it work?
• You need a camera and a marker (reference)
8. How does it work?
• Software analyzes the images from the camera
looking for the marker
9. How does it work?
• Transforms a super-imposed 3D world based
on the orientation of the marker
10. How does it work?
• Camera (to see the world)
• Processor (to analyze the world)
• Display (to show the world)
11. AR Project at KABK den Haag, the Netherlands
Flickr user: przemion
12. Because Mobility is Key
To achieve mobile augmented
reality, we now have:
• Camera with more pixels than
needed
• Large & crisp displays (that are
pocketable)
• Fast Processors
Wikitude on the iPhone
• And now GPS and Compass
enabled! YAY!
13. Augmented Reality in Mobile
• Commercial Apps started popping up around a year ago
• Location-aware devices that are always connected to the
Net opens lots of possibilities
• Most new devices are AR-capable
• But there are obstacles... (and workarounds)
18. Usability and Aesthetic Issues:
• For Info-driven apps, the screen is too small.
• Navigational apps only represent what you can see.
• Games are fun, but the interaction gets tiresome.
22. Locative Art
“... Hollis approached the body. That
wasn’t there. But was. Alberto was
following her with the laptop, careful of
the cable. She felt as if he were holding
his breath. She was holding hers.
The boy seemed birdlike, in death, the
arch of his cheekbone, as she bent
forward, casting its own small shadows.
His hair was very dark. He wore dark
pin-striped trouser and a dark shirt.
“Who?” she asked, finding her breath.
‘River Phoenix’ said Alberto, quietly.”
23.
24.
25. What we set out to do
Placing 3D objects in real spaces.
The objects are tied to a location, but exists in another dimension.
28. How to recognize objects in space?
• Same as conventional AR - we
need a marker
• Feature tracking (OpenCV -
camera vision library)
• Current iPhone SDK allows
developers to overlay graphics
on top of the feed, but not
analyze it.
29. How to recognize objects in space?
• Narrow false-positives using
location sensors
• GPS accuracy is reliable at 30
meters (occasionally 1-3m)
• Works for finding location on a
map, but not pin-pointing exact
spot so that we can “lock” the
objects
35. What is next?
• Hopefully iPhone SDK will open the camera
• Otherwise - give up on the iPhone
• Continue developing recognition algorithms
• Pursue more support to continue
36. But maybe we’re just waiting for magic glasses
Clip from “Realtà Aumentata” by Sorin Voicu
37. But maybe we’re just waiting for magic glasses
Clip from “Realtà Aumentata” by Sorin Voicu
40. Media Credits
Slide #2
Picture of projected map from “Map/Territory” video by timo (on vimeo)
http://www.vimeo.com/5572328
Slide #3
Plug Into The Smart Grid, stevegarfield.com
http://www.youtube.com/watch?v=wVcdurnfcPo
Robert Downey Jr. on Esquire by esquire magazine
http://www.youtube.com/watch?v=wp2z36kKn0s
Slide #24
“Street Tests” composited video by Najork (najork.net)
http://www.vimeo.com/5716181
Slide #33
Realtà Aumentata - The future of education
http://www.vimeo.com/2341387
Notes de l'éditeur
- general intro: hi, we are aesthetec, creative applications of new tech
- artist cred: SMS, bt, brew, j2me, symbian... Footprints
- we are particularly interested in interactions in public spaces
- bringing elements of the online & virtual world into the physical world
- with the support of MEIC, and other industry partners...
- we have been researching augmented reality
- surely most of you have heard of this, AR has been around a long time
- last year was the perfect storm for AR, webcams++, flash was faster, etc..
- GE’s campaign, using Papervision3D in Flash with ARToolkit
- Esquire Mag, remember eInk (Oct’08), barbarian group
- GE’s campaign, using Papervision3D in Flash with ARToolkit
- Esquire Mag, remember eInk (Oct’08), barbarian group
- first you need a camera + marker
- marker is a visual reference that the application knows about beforehand
- a lot of time you’ll see B&W markers, simple
- there are more sophisticated methods that use images (patterns)
- called ‘marker-less’ but is still a marker
- the software looks at the images coming off the camera, and analyzes them for the unique marker
- when it finds the image, it can then detect the 3D properties of that marker: rotation & translation
- finally it uses those properties to orient a graphical 3D world to match those settings, using the marker as the origin
- much like a compass (where north is a magnetic reference)
- essentially it comes down to camera, processor, display
- this system evolved on the desktop, since it could supply the demand
- yet being chained to a desktop is not the optimal to experience AR
- this is a technology that demands to be mobile
- early mobile incarnations of AR looks like this...
- ... because making desktop power portable
- we had to wait for portable devices powerful...
... today we are almost there.
- cameras are more than powerful enough
- displays are large and crisp (as possible for the pocket)
- processors are reaching the 1GHz benchmark
- also added sensors designed to orient people in real space
- accell, GPS, compass
- AR for mobile started popping up a year ago
- Android, WinMo, iPhone, etc...
- devices are location-aware (GPS) and internet connected (3G)
- there are limits, some real/some contrived - cannot access camera
- contrived obstacles demand creative workarounds
- because of the extra sensors you can still orient yourself (without a marker)
two projects coming from europe
- wikitude (mobiliZ): austria
- layar: holland
- acrossair is another company based in europe (London)
- nearest tube
- nearest subway
- screens are too small to see a lot of info,
- brings the problems of real world into virtual. (you have to get closer to see better)
- as a navigational aid it forces you to use it constantly, rather than observe a route like top-down map
- games are fun, but get tiresome after several minutes, camera background is dizzying
- the overall issue is that they aren’t necessary to the experience
- top-down maps, semacode references virtual info, games we want to move less
- novelty vs. necessity?? (better title)
- so... what have we been working on?
- we started talking a lot about AR from spook country
- one main character is a ‘Locative Artist’, creates art works pinned to real locations
- READ
- died from a drug OD on halloween 1993
- what gibson describes is real space frozen in time, saved in virtual space
- then recomposited over real space
- new form of sculpture, whereas moments would be capture in bronze
- imagine the battle of york
- a new way to experience history
- a new way to experience space
- facilitated by technology
- my comfort language is ObjC
- trees are horrible markers, too many variations
- we needed an object that can be tracked
- horizontal lines of the bench
- our next step was to use the features of the bench to place a ball
- sounds simple, right?
- we’ve been working with a student intern from Ryerson University, after a couple weeks of hacking away at this...
- which is what we understand to be the case
- we needed to build more tools
- now we have an environment for testing
- some work-in-progress algorithms