More Core Tools for Augmented Reality

More Core Tools for Augmented Reality

Last week Microsoft held an "Enabling Innovation Through Research" event at its Cambridge research labs, and demoed many of its projects. Core Tools for Augmented Reality was one of those projects, presnted by Simon Winder:



Well, there's nothing in this video we haven't seen before, to be precise, two months ago at Microsoft's Techfest (click for my previous blog post). Even the same bubble oriented treasure hunt game was shown then. The technology itself is based purely on image recognition, using the same concepts behind Microsoft's Photosynth. Some more details (but not too many) can be found on the project's web page.

Via The Future Digital Life and Developement Memo for Ourselves.

Microsoft Demos AR Applications in Techfest 2009

Microsoft Demos AR Applications in Techfest 2009

Some cool news come this week from Microsoft. Apparently, the beloved friendly giant holds its annual Techfest this week, where it demos some of the research projects it has been working on. This year, two augmented reality are featured, NoteScape and Core Tools for Augmented Reality.

NoteScape is a project by Darren Edge, from Microsoft Research Beijing which brings us AR sticky notes. NoteScape lets a user attach virtual sticky notes to her laptop screen so that they follow the user wherever she chooses to go. They can be viewed at any time by using a web cam or a mobile phone camera. More info can be found here, but I don't see why the user can't just use a notepad document (or if she's adventurous, a wiki) to jot down her ideas.



Next comes Core Tools for Augmented Reality, which I guess is a codeword for using Photosynth as a service to drive augmented reality. To the uninitiated, Photosynth is a really cool project by Microsoft, where multiple images of the same object taken from different angles and positions are merged together into a point cloud, indicating where each photo was taken, and provide an almost three-dimension reconstruction of the object. All that, without having meta data about where the images were taken (no compass reading), using only image processing techniques. Here's a good video than explains it better than I could ever could.

Now, imagine a room or an outdoor site, previously "mapped" using Photosynth, for example, Piazza San Marco in Venice. When a user takes another photo at the piazza, Photosynth can quickly tell where the user was located and at what direction he was looking at, with accuracy that can rival civilian GPS devices and compasses, leading the way to augment that picture. That's exactly what Michael Cohen, a principal researcher from Redmond, thought of, as he explains in the next video, where he uses this technology for a virtual indoors treasure hunt:



Only downside? This could probably not run on a cellphone anytime soon, but otherwise, I think its a breath of fresh air, especially when compared to other AR demonstrations we recently encountered.