Image-to-Image Style Enhancement Workflow

Use the attached image as the visual source and the SD1.5 prompt below as the semantic and stylistic source. ORIGINAL PROMPT: Ink illustration on paper, fine line work with opaque flat color, matte print finish, a woman with long black hair in a gothic outfit, black patent leather corset and skirt, cutout sleeve, pale face, dark moody setting, stylized graphic art style TASK Produce a single refined image-to-image enhancement of the attached SD1.5 image: the strongest possible version of the same idea, executed with full commitment to the artistic style the original prompt was reaching for. This is an art-director pass, not a redesign. STEP 1 - PARSE THE ORIGINAL PROMPT Before generating, sort the prompt into three buckets: - SUBJECT: who/what, age, pose, outfit, props, setting, genre, mood, narrative. - STYLE: medium, technique, named artists or movements, render descriptors, lighting language, palette cues, era, format (poster, plate, cel, etching, photo). - NOISE: quality spam and booru filler ("masterpiece", "best quality", "8k", "ultra detailed", "trending on artstation", "highly detailed", weight syntax, negative-prompt residue). Discard the literal tokens but read them as intent - they signal how finished and deliberate the result should feel. STEP 2 - RESOLVE THE STYLE INTO A SINGLE COHERENT IDIOM The STYLE bucket is the aesthetic brief and takes precedence in every rendering decision. - If a physical medium is named (oil, gouache, ink, charcoal, woodcut, screenprint, 35mm film, watercolour), render it with real material behaviour: visible mark-making, brush or nib pressure, paper tooth, ink bleed, grain, halftone, plate registration, edge quality, tonal limits of that medium. Do not deliver smooth generic digital rendering with a texture overlay. - If artists or movements are named, extract their actual formal habits - palette, edge control, line weight, value structure, compositional geometry, level of abstraction, treatment of light - and apply those. Do not merely gesture at a vibe. - If several styles are listed and they conflict, choose the dominant one, subordinate the others as accents, and commit. A confident single idiom beats a hedged blend. - If the prompt names no style at all, infer the most fitting one from the source image's genre and mood, and execute it decisively. - Apply the idiom uniformly across the entire frame. Background, props, atmosphere and subject must share the same technique and level of stylisation - mismatch here is a signature SD1.5 tell. - Let the style dictate the palette, contrast, edge treatment and finish, rather than defaulting to a modern high-saturation, high-bloom digital look. STEP 3 - PRESERVE IDENTITY Keep the core of the original recognisable: main subject, apparent age, pose and attitude, outfit concept, key props, genre, mood, and the central visual idea. Treat the prompt as a statement of intent, not a checklist to satisfy literally. STEP 4 - COMPOSE DELIBERATELY Strengthen focal point, silhouette, gesture, depth, camera angle, framing, negative space and visual hierarchy. Use elegant cropping, clear foreground/midground/background separation, purposeful leading lines, and a single readable idea. Prefer compositional conventions native to the chosen style. Remove clutter that does not serve the scene. STEP 5 - RAISE THE CRAFT Upgrade lighting, colour harmony, atmosphere, texture and material rendering within the idiom. Replace muddy or generic passages with confident value structure, controlled contrast, coherent shadows and an intentional palette. Fix anatomy, hands, eyes, facial structure, clothing logic, object contact, perspective and lighting consistency. Make expressions subtler, sharper and more emotionally readable. Avoid cartoonish distortion unless the style calls for it. CONSTRAINTS Add only detail that reinforces the concept, clarifies the narrative or strengthens the composition. No extra characters, floating objects, irrelevant props, readable text, logos, watermarks, anachronisms, borders or captions. FINAL CHECK Style fully committed and consistent across the frame; medium reads as itself; clear focal point; strong silhouette; coherent lighting; plausible anatomy; clean hands; readable expression; uncluttered staging; no accidental text, extra limbs or stray objects; original concept intact. Output exactly one image.