Tuesday, November 3, 2009
ISMAR09 and the future of AR
12:19
I had a great time at the International Symposium on Mixed and Augmented Reality 2009 (ISMAR09). I met lots of amazing and smart people. It’s great to see the faces and talk to people I’ve been following on blogs, twitter and videos.
I’m an engineer so I like to know what makes AR “tick” but, nowadays I’ve been more interested in the interaction side of AR. The first couple of days I attended the “Science and Technology” sessions but, on the last two days, I decided to check the “Arts, Media and Humanities” sessions.
My first reaction to the “Arts, Media and Humanities” sessions was that they had little to do with AR but, I had a great time. One thing artists do very well is question the system.
I had the same feeling about the keynotes. They were great and presented by very interesting people but, were they related to Azuma’s AR definition? At the first glance, they weren’t:
Mark Mine (Walt Disney Imagineering) – Showed how a make believe world is done. (I had the chance to ride all the attractions of “Future World” at Disney Epcot and it is AWESOME.)
Natasha Tsakos (Up Wake) – Did a performance in synch with projected videos. (I did enjoy the show and had the pleasure to talk to her. She not only did the keynote but also participated in the conference and seamed honestly interested on its technological side.)
Pattie Maes (MIT Media Lab) – Presented the famous “sixth sense” project. (Some people argue that it’s not AR.)
Was the ISMAR09 organization wrong to bring these people or was it like an artist questioning the AR state of the art?
Azuma’s AR definition is correct and I agree with it from the point of view of a computer vision scientist but, is there AR beyond it?
“Sixth sense” involves a projector just like YDreams’ interactive floor projections. In the following video you can see “Virtual Garden”, one of the first applications we created.
It’s arguably AR but, isn’t it “augmenting” these kids world, while they roll on the floor and chase the rabbit? What if instead of projecting on a flat surface, we had projected on real-life volumetric objects? That’s what “sixth sense” and video mapping do.
Although I wrote before about the lack of added value of current AR, I’m still a supporter and I strongly believe it’s an essential step in the long escalator of incremental innovations. I have the same feeling about projectors. What if all surfaces could be converted into computational surfaces? Neil Gershenfeld’s “computer by the square inch” concept can be the end of projectors. YDreams has already partnered with manufacturers of paper, glass, cork and other materials, and is currently researching on how to make these materials interactive.
AR definition also focuses on sight. Can’t we augment reality using the other senses? Hearing, touch, smell and taste. Sight and hearing are the easiest ones to foul but we may in the future be able to foul the others. (I think I smelled those orange groves at Soarin’.)
I think computer vision will still be part of AR but not only for camera pose estimation. Nowadays computer vision allows face recognition, face and gaze tracking, detection of gender, mood, human activity, etc. All these can be used for interaction and it has the big advantage of being non-invasive.
Will ISMAR peer-reviewers accept a broader definition of what is AR or will they stay loyal to Azuma’s definition?
Tags: augmented reality, epcot center, future, international, ISMAR09, MIT media Lab, mixed, symposium

November 4th, 2009 at 10:53
We already have a basic technology to augment hearing! When we use those small, in-ear headphones, that can’t completely block exterior sound, we can hear both.
It is effectively a “hear-through” device, in the same way AR glasses are “see-through”.
Now we just have to make the right sound play in the right time, sort of what “geo-location” does for vision AR.
The technology isn’t mature, so I don’t know if there are any haptic gloves out there that can do this, but the same could be done in terms of “touch-through”. Even with gloves we can feel rough edges and textures - any small vibration in some parts of the glove in addition to the feeling of real-word objects, would be AR touch.
November 9th, 2009 at 2:00
It seems that Nokia is already working on spatial audio which might be led to a cool hear-through technology. Soundwalk has shown location-based audio services which might be expanded to aural AR services. There are also some ISMAR papers on augmented audio published by Japaneses researchers.
I think this audio-based AR has great potential for mobile AR combined with LBS. Vision-based AR is still not mature for general adaptation. There are also many things which can be done without computer vision.