Cloudscape

Pattern recognition could be used to improve the image resolution of photographs to levels actually higher than those of its optical resolution This is how some artists are able to reproduce a photograph in a resolution greater than the original photograph itself, a recent art which has become known as hyperrealism; hyperrealism is achieved mostly through human pattern recognition. In this way, the resolution of an image could be made far greater than the original. This would not be very useful to distinguish additional details, but it could vastly improve the esthetic value of an image.

For example, zooming in on an image until one eventually discerns its pixels, through pattern recognition one could make all pixels in the image transition into one another in a gradient. This gradient, being composed of layers of hues transitioning between one hue and another, could have its layers of hues be equidistant from the nearest border. In other words, each layers would follow the same line or curve, although smaller.

For instance, take a picture of the nose. Zoom in on the nose until you are eventually in its tip, and the pixels become visible. Here, the gradients between the pixels could be aligned with the curve of the tip of the nose, meaning that their layers would be curved in the same way, remaining at an equal distance from the edge. In this it would follow the form of the nose. Because the edge of the nose would usually be marked by an obvious line beyond which there is a sudden change in tint.