오버스케일 팝 스크린프린트 여행 초상화

Please convert the uploaded travel photo of a single person or a small group into a 3:4 vertical "oversized color-blocked silkscreen travel portrait" comparison poster. The image uses a clear upper/lower dual-panel structure. The upper section takes about 42%, preserving the original photo completely and clearly, stretched edge-to-edge horizontally: keep the identity and number of people, face, pose, direction of movement, clothing, carried items, location environment, shooting angle, perspective, natural light, and original color relationships. Only a slight art-publication color grade is allowed. The upper section must remain real photography - no illustration, no replacing people, no changing poses, no mirroring, no blurred edge-filling, and no invented landmarks. The lower section takes about 58%, remaking the same person as a bold hand-pulled silkscreen portrait. Do not replicate the photo's full framing, and do not keep the original distant or full-body scale. Prioritize extracting the face, hair or hat, eyes, mouth, hand gestures, shoulders and neck, collar, and the most recognizable action. Enlarge the figure to 105%-125% of the lower area. Allow the face, brim, hair, arms, shoulders, or clothing to boldly cross the top edge, left/right edges, and bottom edge, so the figure presses down on the entire layout like a close-up view. The figure must touch at least three edges of the frame and must not be shrunk and centered. Facial identity, expression, gaze, hand gestures, clothing silhouette, and main posture must be traceable to the original image. Do not distort the face, change actions, or add limbs for graphic effect. Keep only 1-3 most location-identifying fragments from the environment, such as a building color plane, a section of vehicle, railing, ridge line, wave, ship rail, grassland line, or animal pattern; compress them into a few large color blocks or coarse halftone silhouettes. Do not redraw the entire scene. The color scheme uses warm uncoated paper as the base, adds black structural ink, then extracts 2-4 high-contrast spot colors from the original photo. The palette must change with each photo, allowing warm/cool, light/dark, or complementary color collisions, but fixed orange-blue templates and rainbow palettes are forbidden. The face, clothing, and background are built from large opaque spot-color areas, black outlines, and paper-white negative shapes. Strengthen real print-production marks: very visible coarse halftone dots, uneven ink coverage, dry-squeegee drag streaks, stencil rough edges, pinhole paper-white spots, paper fiber, localized ink skips, and about 2-4 mm registration offset. Misregistration should create clear colored edges at figure outlines and color-block boundaries, but must not destroy facial features or hand gestures. No smooth digital gradients, glossy vector color blocks, generic pop filters, or clean photographic threshold effects. The lower section must have full-bleed, face-close, high-pressure visual impact, while deliberately keeping about 15%-20% of continuous low-information spot-color fields or paper surface, so the text becomes a heavy color block in the composition rather than a small note in the corner. Add only the following text: "[Location or scene English title]" "SCREENPRINT [number]" "[one line of 5-9 English words describing an actual visual observation]" The main title uses huge, narrow, all-caps overprinted lead type, taking about 22%-38% of the lower width, and must be clearly readable in chat preview. The field and observation line are smaller, but must not shrink into gray lines. All text uses the same ink colors from the image, with ink skips, coarse dots, and slight misregistration, as if printed on paper together with the figure; no smooth floating digital fonts. Avoid: small figures, centered full-body shots, same-scale same-composition redraws, full background coverage, generic pop filters, rainbow palettes, overly fine dots, smooth gradients, overly small text, facial distortion, extra people or limbs, irrelevant skylines, author names, artist names, signatures, brands, logos, URLs, watermarks, QR codes, fake Chinese text, and random gibberish.