You know it's a good paper when you reach the end and ask "is that it?" and yet no one else has thought of it before.
Side note: local-global approximation seems to do well in graphics and visual tasks. For example in the field of alpha matting the state of the art for a while was KNN Matting which sampled locally and globally. Most methods since then have taken a similar approach.
Side note: local-global approximation seems to do well in graphics and visual tasks. For example in the field of alpha matting the state of the art for a while was KNN Matting which sampled locally and globally. Most methods since then have taken a similar approach.