[reneberlin]

compare it to jp2a / ascii-image-converter In ascii they perform very well without ML, of course. You could set up a training based on that.

[mntalateyya]

btw I optimized it for edge detection, but it pretty much ignores luminosity. So it performs better at rendering text / line art, but worse at pictures without clear boundaries

[mntalateyya]

Thanks. Very helpful.