Misty / Kasumi カスミ

Multi-step AI-image with FurryToonMix, initial generation with FurryToonMix + h_Madoka's Misty LoRA

Process:
Initial generation with FurryToonMix at 384 x 672 with FurryToonMix & h_Madoka's Misty LoRA v2 (0.6)
Upscaled 2x to 768 x 1344 with sharpness and contrast enhancement.
Then individual sections cropped and redone at higher resolution img2img with just FurryToonMix (no LoRA!) and layered in with GIMP.
Then lots of final touch up work with smudging, redrawing, re-lining, tweaking, shaping, color matching, and shading all by "hand" (or literally my laptop track pad because my mouse broke lol) etc
Then upscaled again, sharpened and artifact removal, descaled back to original size and cut up and parts lightly blended with the prior image for just the right look, then even more re-lining, shading, and pixel edits etc until done to perfection... *chef's kiss*

Not a one-shot generation, but a nice exercise in enhancing an initial image, mostly using FurryToonMix (SD 1.5 model) for the major sections.

If you like the result buy me a drawing pad and an awesome NVIDIA video card and I'll make perfect images all day long.