3 ms·
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.
by reneberlin 10mo ago
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 10mo agoThanks. Very helpful.
- mntalateyya 10mo agobtw 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