[kaahne]

Yes and no. Going through the paper real quick, we use similar technique, but do not analyze the video to reconstruct the movement of the camera, which allows us to speed up the algorithm significantly.