One of the most popular examples of using genetic algorithms in image processing is the reconstruction of a given image with a set of semi-transparent, overlapping shapes. Besides the fun aspect and the opportunity to gain experience in image processing, these experiments provide an excellent visual insight into the evolutionary process and could potentially lead to a better understanding of visual arts, as well as developments in image analysis and image compression.
In these image reconstruction experiments – multiple variations of which can be found on the internet – a familiar image, often a famous painting or a segment of it, is used as a reference. The goal is to construct a similar image by assembling a collection of overlapping shapes, typically polygons, of varying colors and transparencies.
Here, we will address this...
