Panoramic views and the outer space - visualizations at Oredev

So, it's November 9th 2009 and the Oredev conference has taken off; for me, it's 3 days of talks about the software industry on topics that I most enjoy. In the spirit of the times, I kicked off my first day with a popular topic - visualization.

Eric Stollnitz from Microsoft showed that there are more cool things to MicroSoft than the XMLHTTPRequest object. We were shown a slew of demos dealing with images and image processing. The common theme running through these applications was taking disparate chunks of visual information (images) and transforming them into a form natural for the human perception to process. There was an application that, given a large set of photographs, stitched them together into into one complete, continuous view. Imagine having taken a number of photos (of a landscape, for instance) from different angles. Navigating these images one by one has been the domain of a traditional photo gallery; the demo showed how these discrete photographs can be viewed much like the original, real-life original. To this end, images are "stitched" together and 3D panoramic effects are applied to the resulting landscape to make the viewing close to viewing the original. A flat photograph acquired depth and volume; zooming in felt like stepping up closer to the picture. (An aside: I never knew there were owls in the daytime on the rooftops of Seattle!)

The computational challenges of solving such a problem efficiently are interesting. For example, when handed N images (or pieces of a puzzle), how would you put them together to the original form without comparing every image to every other image (N*N comparisons)? Leave a comment below.

Other demos included The Worldwide Telescope, a collaboration between MicroSoft and various astronomy labs. Viewing planets and moons in real-time, and even zooming out to view or Galaxy (which looks like a fine, white mesh, like a coral reef suspended somewhere in the Universe) has never been easier.

All in all, this was an interesting presentation offering one possible answer to the question: "How do we best use the enormous computing power we have at our disposal"?

Eric's answer, with this presentation, is: "Help us visualize data to understand it better".