Yeah, Ryumaru explains it better. It's about distribution of details. How many areas are there with high detail, how many with low, medium, etc. If you go with a gauss distribution or golden ratio or similar, you end up with a similar visual distribution to the Sierpinski triangle, so it's a great reference for stuff like this. You'll find that people like Da Vinci were heavy into this stuff. To see it in the mega man screen, you need to squint, so it blurs a little. If you squint, everything is still recognizable, a lot of detail is still visible, etc. If I squint on yours, the skulls become blobs (since they only contain high detail), the rest becomes big swats of color and the bread stays just the same (aka, there's probably some medium detail lacking in it).
High detail is something like AA or single pixel outlines (the skull is almost fully high detail).
Medium detail is something like shading or the facial features on the bread or the text.
Low detail is just the big swats of color, like for example the square shapes.