We’ve all seen it. Pick your favorite crime fighting drama on TV and inevitably the cops will take some grainy wide shot from a random security camera, then run it through their magic zoom and enhance procedures to reveal a clear shot of a license plate parked several blocks away. Real methods do exist that are capable of some image enhancement, but unfortunately, in the real world, those cop drama techniques are simply Hollywood special effects.
Cameras have been increasing resolutions (2K, 4K, 8K, etc.) which helps the smaller details of an image become clearer. However, even with the larger resolution cameras, you simply cannot magnify the image without some loss in image quality.
Researchers from the University of Bath in the UK, have been attempting to solve this problem with a technique that graphic designers and animators have been using for years. By applying vector graphic technology to a video image, one could theoretically zoom into any portion of the screen and still maintain perfect image quality. Take a look at the graphic example below:
There are three parts to this image. The top portion shows the entire logo for Sundog’s podcast. The lower portion zooms into the word “THE,” with the left half showing the result of magnifying the logo using vector technology, while the right half shows the result using raster or bitmap magnification. The difference is very clear, or fuzzy depending on how you look at it. This logo and accompanying artwork were designed using vector-based designing software. Vector graphic are mathematical expressions of points, lines, curves and polygons. Each point on a vector graphic has a specific X and Y position. The relationship of these points do not change when increasing the size of the artwork. Everything stays nice and clean. On the other hand, raster images are pixel-based, and cannot be magnified without a loss in clarity.
The University of Bath wants to develop a new method of capturing video that uses vector technology over pixels. This report says “…4K may become obsolete along with resolution itself—killed not by 8K or Super-Hi-Vision but by a completely different kind of technology. Keep in mind that this technology has a long, long way to go to become practical. Remember, I said vectors are based in mathematical expressions. It’s going to take a lot of computing power to translate a single video image into a vector-based data file. But as long as Moore’s Law continues to exist and computers double in speed every couple of years, perhaps that magical zoom and enhance button isn’t far behind.