By interpreting the sketch drawn by the user in light of 3D information obtained from computer vision techniques, a small number of simple 2D interactions can be used to generate a realistic 3D model.
In many ways the system is similar to the sadly demised CANOMA but with video, the movie below provides a good insight into how Video Trace works:
By way of simple comparison with CANOMA below is a movie we created some 8 years ago from a postcard - if we had the video capability it would of provided a similar output to Video Trace:
A version of CANOMA but from video is an intriguing step forward - if this can be semi-automated then low end rapid capture of city models could be round the corner.
You can find out more about Video Trace via their main page.