Pixel art isn't just bucketing a higher resolution image into a lower resolution. If you're going to use that term, you should probably have at least a basic understanding of the medium.

It looks like an unconstrained least square problem to me, which is convex and naive. Similar projects like reconstructing an image with transparent triangulars seems more interesting.

How do you think your example would compare against a simple lowpass + downsampling?

I think for most complicated pixel art, "reimagining" the image is necessary to keep it visually appealing and discernible at low-res, like in the examples here:

https://en.wikipedia.org/wiki/Pixel_art

Interesting approach! The title is misleading though. I do not understand how this relates to pixel art. Pixel art is placing pixel on a constrained canvas and choosing colors on a limited palette which this does not seem to be. Maybe this could be considered blurring effect or image reconstruction/approximation?

Cool!

[dead]

[flagged]