103내 사진 3D 애니 장면

Generate this image now. Draw it directly - do not describe it, explain it, or ask questions.
Output size 2480x3508 pixels, A4 portrait, 300dpi, print-ready poster.

[Purpose]
Create exactly ONE complete portrait image that turns the attached personal photo into a single frame from an ORIGINAL warm theatrical 3D animated film. The owner will share this one image on Instagram or Threads. The result is one continuous scene on one canvas, not a storyboard, comic, film strip, phone mockup, poster advertising a film, before/after comparison, or a set of variations. No cropping, code, further turn, or assembly is required. The fixed result is a polished animated scene; do not offer alternative art styles.

[Scene]
Start with what is really visible in the attached photo. Preserve the number of people, their relationship as visible, clothing silhouettes, body positions, recognizable objects and the setting. Convert them into stylized but recognizable three-dimensional animated characters with expressive eyes, rounded yet individual facial geometry, natural hand anatomy, soft skin shading, and believable hair volume. Preserve the specific memory "【사진 속 기억】" only as context; do not illustrate an unphotographed event merely because the text suggests it. Pay particular attention to "【꼭 살릴 것】". If the supplied detail conflicts with the photo, use the photo as the visual source and keep the detail only where it truly appears. The moment should feel candid and alive, as if captured between actions. One gentle visual narrative beat is enough: a shared look, mid-step motion, or a small reaction already supported by the original pose. Keep the same viewpoint and approximate framing.

[Attached photo]
The user attaches ONE photo in the SAME message as this prompt. Use it as the identity and scene reference. Keep each person's apparent age, skin tone, face shape, hairstyle and glasses recognizable. Do not beautify into a stranger, swap faces, change the number of people, add a child or companion, or invent a costume. A photo without people is unsuitable for this portrait card; request a photo showing at least one person instead of fabricating a protagonist. Do not treat a background passerby as a new lead.

[Text layout]
Place one short Korean scene title "【짧은 장면 제목】" in the quiet area around 68–76% of the portrait, away from eyes and hands. If the value is "알아서 골라줘", invent one original Korean title of at most eight syllables that fits only what the photo shows; never print the instruction phrase. The title is the only visible text. Render it once in clear, gently hand-lettered type with a solid subtle shadow. The top 12% and bottom 15% must contain no important letters because the site's listing overlays those regions. There is no film logo, credit block, date, subtitle, episode number or fake studio label.

[Style]
Use one consistent, original feature-animation 3D look: sculpted human proportions with slightly amplified expressions, refined material response, clean forms, rich but restrained color, shallow depth of field and soft cinematic bounce light. Let the attached photo's time of day and weather determine the palette. Rain may give warm window reflections; sunlight may give diffuse golden edges. Do not impose a generic sunset over every image. Faces retain small real asymmetries, and clothing folds follow the actual posture. The environment is a coherent modeled version of the original place, not a newly invented fantasy world.

[Visual hierarchy]
At thumbnail size, the viewer should first recognize the animated person or people and their interaction, then read the small title. Keep heads large enough to see and leave breathing room around them. Avoid stickers, decorative frames and scattered microtext. If there are two people, give both clear faces without reducing either to a background duplicate. Keep all figures in one spatially continuous shot, with shadows and eye lines that agree.

[Lighting and texture]
Use physically plausible contact shadows, gentle surface roughness and clean volumetric light only where a real light source could create it. Text remains sharp. Avoid plastic shine, waxy skin, heavy glitter, oversaturated neon, artificial lens stars, repeated windows and warped geometry. The 3D finish should look cinematic on a phone rather than like a toy in a box.

[Text rules]
Proofread every Korean syllable. Print the supplied title exactly once, or invent only the title if the input requests it. Never print field names, braces, "알아서 골라줘", the memory note, the preserved-detail note, English filler or watermarks.

[Avoid]
No named work, artist, animation studio or copyrighted character. No four panels, multiple frames, split view, figurine, package, wallpaper clock area, cosplay photograph, extra limbs, changed identity, invented landmark, false event or explanation outside the image. Return only the finished image.