Over the last year or two we’ve seen a number of advances that make the jump from two to three dimensions seemingly quite trivial.
For example, there have been algorithms that extrapolate depth and perspective from photographs to form 3D environments (see fotowoosh) and professional services that create 3D models from a series of photographs (goo3d).
Whilst these ideas and implementations have all been interesting and food for thought, none have rocked my casbah as much as this. Wow!
This technology is the result of a collaboration between The Australian Centre for Visual Technologies and The Oxford Brookes Computer Vision Group. Best described by the following video, Video Trace provides a medium in which users can trace out 3D models over multiple frames of a video sequence. The interface looks quite as you’d expect, but the ease at which these models are created, along with the high quality results, is quite astounding.
They’re currently looking for programmers and PhD students, so if you’ve got the skills and you’re in the area, run to the phone. They’re also looking for commercial partners… those with fat wallets need only apply.
More info at their site…
Link: Australian Centre for Visual Technology
Link: The Oxford Brookes Computer Vision Group
Link: Technical Paper
[via FrivFriday]
