Korean Naive Flat Editorial Poster

Korean Naive Flat Editorial Poster

Use the system imagegen skill and image-generation tool. Treat every uploaded photograph as the edit source. Preserve its layout and story while simplifying form, color, and line into a deliberately designed naive illustration.

Output contract

• Produce one independent raster poster for every uploaded photo. Never combine photos into a collage, contact sheet, comparison layout, diptych, or shared canvas.
• Use a 3:4 portrait canvas with one exact horizontal 50/50 division. Keep truthful photography in the upper half and the naive flat illustration in the lower half. Do not let either medium invade the other.
• Preserve immediate correspondence through the hero's identity, posture, proportion, direction, main relationship, and approximate source layout.
• Keep one clear visual center, a small cast of supporting forms, flat color, and controlled imperfect black contours.
• Render visible wording in English unless the user explicitly requests another language.
• Return every poster as a separate image. Keep user-facing commentary to one short status sentence only when necessary.

Preserve the photographic upper half

Keep the source recognizably photographic. Preserve subject identity and count, anatomy or construction, proportions, pose, silhouette, direction, spatial relationships, real materials, natural light, and the original color atmosphere.

Apply only restrained art-magazine, independent-publication, or exhibition-photography color correction. Do not redraw, vectorize, beautify, materially relight, or apply cartoon outlines, flat color, typography, or graphic overlays to this region.

Crop deliberately for the upper half. When more area is needed, extend only plausible sky, ground, wall, water, foliage, or environmental background. Never stretch, warp, mirror, duplicate, or redesign the subject.

Extract the simple story

Before prompting, privately identify:

1. the one hero subject or inseparable relationship carrying recognition;
2. its decisive silhouette, posture, proportion, action, direction, and facial or structural cue;
3. three to six identity features that must survive exaggeration;
4. the approximate placement, object-to-object relation, and core action from the source;
5. the supporting objects or environment shapes required to explain scale, place, or story;
6. the visual center of gravity and useful size differences, offsets, overlaps, and crops;
7. the upper photograph's most lively and recognizable colors;
8. one concise English title and only the few small notes that add wit or rhythm.

Delete realistic lighting, perspective detail, material texture, background clutter, and small description. Do not expose this analysis or repeat the user's design brief.

Simplify with rounded, slightly awkward geometry

Translate the hero and essential supports into round, simple, lightly exaggerated shapes. Preserve the core gesture and relationship while allowing a slightly oversized head, simplified body, broad leaf, compact building, enlarged wheel, softened corner, or other source-responsive exaggeration.

Use a few broad masses and clear overlaps. Keep limbs, branches, roofs, vehicles, objects, and terrain easy to read. Let small irregularity create personality without damaging anatomy or construction.

Do not force the same cute proportions on every subject. Buildings, landscapes, food, vehicles, and objects should retain their own identity rather than become character mascots. Avoid chibi formulas, anime faces, emoji expressions, bubble-headed people, generic smiling animals, and sticker-like white borders.

Draw imperfect contours with discipline

Use black hand-drawn outlines with moderate width and visible but controlled imperfection:

• small pressure and width changes;
• slight wobble on long contours;
• short local breaks at overlaps or turns;
• occasional imperfect joins or offset closures;
• a few blunt starts and tapered ends;
• sparse corrective line echoes only where useful.

Keep contours clear enough to organize flat color. Use very few internal lines for faces, joints, folds, windows, branches, or structural cues.

Avoid smooth vector uniformity, heavy marker outlines, dense scratchiness, repeated hairy tracing, uncontrolled scribbles, grunge noise, photorealistic hatching, and deliberately crude child drawing.

Compose one playful center

Keep one clear hero. Place supporting forms around it through size contrast, offset, overlap, partial crop, and incomplete symmetry. Preserve the source's key relational layout while simplifying perspective into readable stacking.

Use a large bright background field and visible blank space. A local complete color block may occupy one side or section to support the composition, but it must not become a generic template panel.

Build a light rhythm from one dominant mass, one or two medium supports, and a few small details. Avoid evenly spaced objects, repeated icons, complete scenic backgrounds, crowds, many equal characters, decorative confetti, borders, and multiple focal points.

Present the result as a flat editorial artwork. Do not add a frame, wall, notebook, tape, stickers, desk, picture-book mockup, product packaging, or commercial card layout.

Translate a bright limited palette

Extract the upper photograph's most lively and recognizable color relationship, then raise lightness and purity while limiting the number of flat fills. Use roughly three to six active colors including the background and black contour.

Bright blue, red, yellow, green, and white may appear only when supported by the source. Preserve its warm/cool identity and assign colors by mass and hierarchy.

Use a large background color, one or two principal hero colors, a few supporting fills, black contours, and at most one small accent. Keep surfaces matte and flat.

Avoid fixed primary-color cards, gradients, realistic shadows, gloss, muddy gray, complex mixed color, neon, rainbow palettes, candy overload, plastic color, and equal-area multicolor distribution.

Integrate playful English type

Derive one concise English title from the source's subject, action, mood, or story. Add only a few tiny phrases, object words, supported place terms, numbers, or playful annotations.

Use a slightly handwritten, uneven, or gently tilted display style for the title. Keep microtext clear and restrained. Arrange wording along a hero contour, color-block edge, blank area, or visual axis. It may rotate, offset, split, or lightly interweave with the illustration.

Make text feel drawn with the image rather than pasted afterward. Avoid a commercial headline, long copy, forced jokes, unsupported locations or numbers, issue codes, logos, brands, signatures, pseudo-language, or watermarks.

Compile the image-edit prompt

Write one source-specific production prompt per input photo containing:

1. one independent 3:4 portrait output and the exact 50% photography / 50% naive-flat split;
2. the upper-half preservation invariants, restrained color correction, crop, and any safe background extension;
3. the lower hero, relationship, source layout, posture, proportion, action, direction, and three to six identity features;
4. the exact rounded geometric masses, source-responsive exaggerations, simplified overlaps, supporting forms, and deleted realistic detail;
5. the imperfect contour width, wobble, local breaks, joins, endpoints, minimal internal lines, and controlled correction marks;
6. the visual center, size hierarchy, offsets, crops, incomplete symmetry, supporting color block, blank-space shape, and reading path;
7. the exact three-to-six source-derived active colors, mass assignments, black contour, matte fills, and contrast hierarchy;
8. the exact English title and tiny annotations with their contour-, block-edge-, blank-space-, axis-, rotation-, offset-, or split-aware placement;
9. hard exclusions: no multi-photo layout, no altered or redrawn upper subject, no approximate split, no photo cartoon filter, no chibi or anime, no generic mascot face, no white sticker border, no smooth vector line, no uncontrolled scribble, no realistic shadow, no gradient, no 3D, no dense scene, no fixed primary palette, no commercial card, no logo, and no watermark.

Use concrete source layout, shape simplification, line imperfection, color assignment, and type placement rather than relying on adjectives such as “premium,” “Korean,” “cute,” or “naive.”

Inspect and correct

Review every result at full size and thumbnail size. Verify the exact split, truthful upper photography, immediate lower correspondence, one hero, preserved action and layout, rounded simple forms, controlled imperfect contour, bright limited palette, large background/blank space, and integrated English type.

If it looks like a generic cartoon, restore source layout, posture, and identity cues. If it becomes too polished, add restrained wobble and local contour breaks. If it becomes childish or messy, simplify exaggeration and remove extra line noise. If color feels fixed or commercial, reconnect each fill to an upper-photo hue. If supporting objects compete, reduce their scale and count. Regenerate with one targeted correction when needed.