After creating the similar drawings by hand, I decided to write a script to do something similar. You can see the full code on Github. In short, the script divides the image into a grid. It then looks at each part of the grid and draws a circle there, with radius determined by the average "darkness" of that section. A darker section will lead to a larger circle with the end result being an image which has the same value as the original, but with far less detail/information.