I know I'm requesting a lot, but I hope the dev likes my ideas so the tool can be super great.
Another idea I have to realy improve the "pixel art" thing, since I'm mostly an artist, would be something like a grid, similar to a Lattice modifier in blender or just Warp (I think is called warp in photoshop).
The point would be that a basic grid of vertex points could be used to make some feature, like the head or eye, bigger on really low resolutions when downscaling, like increasing the density/importance for the algorithm to keep details.
It's an issue I see right now with pixel over, that when you downscale, some details like eyes just get downscaled too much, when they should become bigger proportionally when downscaling (as a real pixel artists would do).
This would allow to make the head bigger and eyes bigger into some chibi SD anime character, which would allow to easily make 32x32 or 64x64 sprites with no issue. (Most algorithms to downscale simply keep everything proportional, when a real human would simply make the head bigger and the eyes bigger when downscaling).
Another way to do this, now that I think off, instead of a grid, would be simply to make a rectangular selection (you could make multiple "selections") to transform a detail in the drawing to become bigger (by example, two "selections" of each eye and some hair bangs) when downscaling to 32x32.