Reconstructing images with polygons

One of the most captivating applications of genetic algorithms in image processing is reconstructing a given image using a collection of semi-transparent, overlapping shapes. This approach is not only enjoyable and educational in terms of image processing experience but also offers a compelling visual representation of the evolutionary process. Furthermore, these experiments can lead to a deeper understanding of visual arts and potentially contribute to advancements in image analysis and 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 fragment 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 challenge by utilizing the genetic algorithms approach and the deap...