My original rendition of strange matches with my current level of skill with digital art didn't really make the impact I'd like, and thats fine. It just means I need to continue working to develop my own abilities, studying, watching tutorials and practice practice practice.
But I figured I would at least play with Stable Diffusion, and try to crank out what I was hoping to bring about~<3 As always I will try to highlight what is assisted with AI just as I highlight in my gallery what is simply a recolor or fanart <3
So Image 4 is my Original, and from right to left is me steadily trying to scale up the detail and horror aspects with the help of AI.
Graviti Diffusion AI Art Generator Model: Domesticated Mix
Primary Prompt: Furry, Anthro, mouse, white fur, white colored fur, blonde hair, blonde colored hair, red scarf, green plaid skirt, ((orange sweater)), holding torch, torch, blue fire, blue colored fire, library, bookshelf, books, wooden floor, spider web, suit of armor, blue eyes, fear, scared nervous, shy, cute, shock, ((Resident Evil 4)), Ashley, tiny, small, smaller female, medium breasts, small female, large room, size difference, huge room, paper, brown belt, belt buckle, paws, boots, walking, clenched hand, followed by monster, monster, monster in shadow, monster on background, Las plagas, followed by suit of armor, followed by Las plagas, mist, fog, scary, cold, monster with glowing red eyes, match, ((holding match)), matchbox, blue matchbox, reflective floor, dusty floor, ((pink mouse tail)), followed by bug, followed by creature, followed by bug creature, all fours, crawling, knight, followed by knight, sleeveless
masterpiece, high quality, best quality, sharp focus, detailed fur, detailed background, detailed eyes, shimmering eyes, sparkling eyes, detailed hair, smooth, depth of field, high res, 4k, 8k, fantasy, horror, shadows, spider web, dust, dust particles
<lora:Furtastic_Detailer:0.5>
Negative prompt: ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, bad anatomy, watermark, signature, cut off, low contrast, underexposed, overexposed, bad art, beginner, amateur, distorted face, blurry, draft, grainy, badhandv4, bwu, dfc, easynegative, ubbp, updn,
So it was mostly iterating from img2img, but since I was trying to preserve as much of original as possible, I push my original illustration into control net again strictly as a reference at a low influce,, thats why img three came out much flatter, which I used later without control net and used a combination of img2img playing with the furtastic detailer and some inpainting to maintain the monster lurking behind.
I got alot of interesting variations thru this whole process ^w^ ♡ I do wanna mess with controlnet open pose more eventually so less is up to random, i am just hesitant
So it was mostly iterating from img2img, but since I was trying to preserve as much of original as
