Photosynth provides new photo experience
Microsoft is promising a blow-your-mind way to view photographs in the three-dimensional context in which they were taken. The technology behind Photosynth includes computer-vision algorithms developed by University of Washington and Microsoft researchers, as well as software for displaying and navigating images built by a Ballard startup company Microsoft acquired.
HOW IT WORKS:
This Photosynth environment starts with a collection of photos taken from different angles and perspectives of St. Mark’s Square in Venice, Italy. Here, each photo is depicted in its own plane, inside a 3-D model. The orange cones represent the position of the camera when the photo was taken. The algorithms automatically discern features in each photograph, such as a door or roofline. When the same feature is discovered in at least two photographs, a point is created. The points accumulate to form a rudimentary 3-D model in which the photos are arranged. The web of relationships among the photos allows the software With the environment in place, the user can smoothly navigate among neighboring images, zoom in and out or explore the space in other ways. You can also view images based on features they have in common. If the mosaic over the entrance to the basilica catches your attention… …you can zoom in for a closer look at all of the images that include it.
A prototype is due out in the fall. Meanwhile, video demonstrations can be found at: labs.live.com/photosynth
The man with the vision Blaise Agüera y Arcas, whose work at the startup company underlies much of Photosynth, said he was always asked about his “elevator pitch” – the one-minute speech on what you’re all about – when he founded the company. “I just hated those kind of questions,” he said. “It’s so intensely visual, this work. There is no elevator pitch. You need to see it.”
THE SEATTLE TIMES