
원본 문서
GPT Image 2 · en
Scene:
A square, straight-on view of a crowded but orderly railway platform during a fictional spring travel rush. The image is composed like a centered stage set: the platform roof, tile grid, bench rows, luggage stacks, and train doors all align to a strict middle axis. Use clear winter daylight, faded teal beams, peach walls, maroon luggage, cream tiles, butter yellow ticket barriers, and soft red lantern shapes with no text.
Subject:
A family of three stands exactly in the center facing the camera: grandparent, parent, and child arranged in a compact vertical triangle, each holding simple parcels tied with string. On both sides, travelers become balanced repeating groups rather than chaos: matching thermos bottles, checked bags, quilted coats, paper snack boxes, folded blankets, and two train doors behind them. The train is present as a pastel green carriage with blank destination panels and rounded windows.
Important details:
The visual language must be controlled through geometry and production design: central frontal gaze, bilateral symmetry, deadpan stillness, miniature-like props, pastel color blocking, and carefully repeated luggage motifs. The travel-rush idea should be readable from density, winter clothing, family grouping, platform architecture, and parcels, but the composition must not become documentary clutter. All signage, tickets, train panels, package labels, and lanterns must remain blank or abstract shapes.
Use case:
Wes Anderson Language single study 6/8, independent transfer into a spring travel railway platform.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no celebrity likeness, no named film references, no real railway branding, no unsafe crowd crush, no photojournalistic chaos, no diagonal station perspective, no extra typography.사용 방법
- 매개변수를 포함한 전체 원본 프롬프트를 복사하세요.
- 주제, 문구, 구도를 자신의 요구에 맞게 바꾸세요.
- 호환되는 외부 도구에서 사용하고 이용 조건과 결과를 확인하세요.
시작 전 확인
- 참조 이미지가 필요한지 확인하세요. 호환 표시는 해당 모델로 이 예시를 생성했다는 뜻이 아닙니다.
- 이름, 수치, 이미지 속 문구를 확인하세요. 모델 버전에 따라 결과가 다를 수 있습니다.





