Our approach trades accuracy for speed, suitable for applications where some accuracy can be sacrificed to gain speedups in the image enhancement.
The most notable such application is that of real-time image enhancement, where the user of a microscope (or other optical imaging instrument) desires real-time feedback on the microscope's subject as it is being scanned. Here, an algorithm's main priority is to be fast, to give the user feedback as soon as possible, without spending a lot of time restoring the image. Once the user is satisfied with this preliminary scan (and the fast image enhancement associated) they may then scan the subject at the highest resolution and possibly with an exact image enhancement algorithm, knowing that while it may take some time, they have an idea of what the result should be like (in terms of quality).1