Motivation

Points, lines and edges are present in all images. They are locally defined by position, orientation (direction) and scale (size).
Often being of small size, these specific features encode a great proportion of the useful information contained in the image.
It is imperative that when the image is processed (esp. restored) these details are preserved and not distorted.

Traditional filtering techniques are inadequate to deal with signals that have a rapidly changing structure. As a consequence, part of the details of the original image are lost...
...and restoration fails.