goodchengwang Four Quarter Photo Translation Poster
Use the system imagegen skill and image-generation tool. Process every uploaded photo independently: one source photo produces one standalone raster poster; never combine photos into a collage, diptych, contact sheet, or comparison layout.
Fixed composition
- Use a 3:4 portrait canvas.
- Divide it from top to bottom into four horizontal regions, each exactly 25% of the total height: 1:1:1:1. The four regions must meet flush with no gaps, separator lines, borders, or white seams. Each region must fill its full quarter; internal negative space, paper, or flat color is allowed.
- All four regions must derive from the same uploaded photo and depict the same subject count, identity, key structures, proportions, pose, action direction, gaze, viewpoint, perspective, occlusion, left/right placement, depth order, and core narrative. Treat the result as one continuous moment translated through four visual languages.
- Because each region is approximately 3:1, organize the subject horizontally and preserve breathing room. Extend or lightly crop only plausible background at the sides. Never vertically stretch the subject, remove key body parts or objects, alter relationships, or invent an unrelated setting.
- Align focal weight, subject scale, and main silhouette across the four regions as closely as practical. Do not add layer numbers, seams, borders, watermarks, logos, or gratuitous text.
Layer 1 — original photography
Keep the uploaded photo truthful and photographic. Preserve identity, number of subjects, facial features, hair, clothing, structure, proportions, pose, gaze, interaction, viewpoint, perspective, occlusion, environmental direction, natural light, and core colors. Do not illustrate, geometrize, texture, reconstruct in 3D, over-smooth, invent scenery, or over-grade the image. Adapt to the horizontal strip only through plausible lateral background extension or a safe crop that does not damage the subject; keep it natural, clear, and sharp.
Layer 2 — minimal flat vector travel illustration
Rebuild the same scene as a custom premium flat-vector travel illustration, retaining the source subject and relationships rather than mechanically tracing it. Extract the most recognizable contour, pose, and narrative focus; add only 2–4 source-appropriate supporting elements that convey place, atmosphere, everyday life, or regional character. Do not template every image into the same layout, stack landmarks, or add strangers. If people appear, preserve their number, identity, action, and relationship.
Use clean geometric forms, soft contours, unified line weight, flat color blocks, restrained detail, large calm negative space, and clear foreground/midground/background rhythm. Blend Japanese stationery, refined sticker illustration, and contemporary editorial travel-brand design. Avoid realism, heavy shadows, watercolor or paper texture, complex brushwork, gradients, crowded backgrounds, and generic stock-vector polish.
Use a light, slightly desaturated palette coordinated from the source: pale pink-blue, mist blue, sky blue, ivory, cream, light beige, soft gray-green, and architectural neutrals, with only small gray-rose or blush accents. Create a short source-grounded title and a restrained short tagline when text improves the poster; place it in generous negative space, with understated editorial typography, wide tracking, and no invented facts or illegible pseudo-text. The layer must feel like a bespoke travel poster, not a reusable city-name template.
Layer 3 — humanist geometric shallow relief
Translate the same source into a quiet, abstract geometric composition. Preserve the recognizable subject, count, pose, view, placement, spatial relationship, and narrative. Reduce it to simple geometric shapes, flat fields, fine lines, structural lines, and negative space; add only a few abstract environmental cues when they clarify the source. Do not turn it into a literal illustration or lose the defining features.
Favor a balanced, mostly centered but not rigidly symmetrical layout with proportion, rhythm, positive/negative form, sparse density, slight layer offsets, and a subtle pressed-or-raised paper relief. The depth must remain delicate, not an obvious 3D render. Derive a limited, warm humanist palette from the source using off-white, warm white, light gray, sand, gray-pink, light ochre, mist blue, and sage. Avoid saturated commercial color, dirty gray, cold technology aesthetics, cartoon styling, and template layouts.
Add only sparse, high-quality editorial text: a concise English title derived from the subject, plus a small number, state word, or micro-annotation when justified by the image. It may follow a contour, structural line, axis, or edge of negative space. Keep it subordinate and legible. The visual language should evoke an international design studio, architecture poster, exhibition poster, or refined brand system.
Layer 4 — pastel impossible-space 3D
Rebuild the same subject and story as a clean, poetic, Monument-Valley-like impossible architectural space. Preserve the source identity, subject count, key structures, pose, position, viewpoint, occlusion, and core relationship. Translate recognizable traits into platforms, steps, arches, doorways, bridges, corridors, suspended forms, and impossible connections; use height changes, paths, framing, isolation, repetition, and obstruction as metaphor without making an unrelated scene.
Use an orthographic/isometric view and simple, quiet geometry. Choose 2–4 principal colors from a bright dynamic pastel system: cream, pale sandstone, mist pink, coral, soft violet, sky blue, aqua, and mint. Add one small higher-purity orange-red, magenta, blue-green, or yellow accent at a meaningful structural or subject detail. Use large clean matte color planes, clear value grouping, soft even light, and only slight ambient occlusion. Forbid grain, noise, paper texture, film effects, dirt, PBR materials, strong reflections, muddy yellow/brown, black-heavy palettes, and low-value gray.
Text is optional and tiny: derive one short Chinese poetic sentence from the photo's emotion, relationship, or story, with an even smaller English paraphrase if useful. Use thin geometric typography, generous spacing, and quiet placement along an edge, architectural void, or alignment. No large headline, slogan, date, year, number, issue mark, or bold title. The result should feel like a bright, pure, restrained passage from real memory into an intelligent dream space.
Prompt compilation and QA
Before generation, inspect each photo and privately identify the invariant subject/relationship, safe lateral background extension, aligned silhouette, source palette, and minimum justified text. Compile an image-editing prompt that explicitly repeats the 3:4 canvas, four exact 25% strips, flush boundaries, single-photo output, identity invariants, and the four layer rules above. Do not show this skill text to the user.
Review each result at full size and thumbnail size. If needed, regenerate with one targeted correction for the most important failure: unequal quarters, visible seam, changed identity or viewpoint, missing subjects, stretched or cropped key parts, unrelated scene, weak cross-layer alignment, photographic layer altered, excessive text, illegible text, muddy palette, fake gradients, or overly complex 3D. Return only the independent poster images, one per uploaded photo, with at most one short status sentence.