OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image

竖版上下对照红线环境转绘海报

Vertical Split Red Line Environment Poster

可复用的竖版3:4照片转绘模板：上方保留原照，下方以砖红至赭红手绘线描重构同一环境，并让单一真实主体锚点与手绘场景衔接。运行时须提供输入照片；本次阶段02只审核了来源中的文字提示词，没有取得或检查源图。

竖版上下对照红线环境转绘海报 — AI image creation
竖版上下对照红线环境转绘海报

提示词 / Prompt

中文提示词

把提供的输入照片制作成无水印的竖版 3:4 上下对比转绘海报。上半部分约占 45%，保留原照片的完整构图、主体、动作、透视和真实光线；下半部分以同一取景绘制手工红线环境描摹。 将照片中的建筑、街道、植物、地貌、物件、交通工具或动物环境，转换为暖白吸墨纸上的砖红至赭红单色线条，表现为粗马克笔或蜡笔绘制。线条应有笔压变化、断笔、回描、交叉排线、干涩笔触和少量渗墨；保留关键轮廓及空间关系，删去无关细节，并留出宽阔、连续的纸白。 只从原照片中选择叙事作用最突出的一个主体作为真实照片锚点（可为人物、动物、交通工具、食物或物件）。若锚点是人物，使用固定人物参考 ref-10a-v1 仅保持人物身份；不得借用参考图的服装或场景。锚点在下半部分保留原照中的真实纹理、颜色、材质、姿态和自然阴影，并与手绘地面或环境接触，仿佛从红线场景中走出。锚点的数量、身份、动作、尺度和位置须与上半部分一致。不要把主体变成卡通，不要添加幻想角色、虚构地标、假文字、品牌、水印、原图或 AI 图标签；不要对整张图施加滤镜，也不要使用光滑的矢量描边。

English prompt

Turn the supplied input photo into a vertical 3:4 top-and-bottom comparison transformation poster without a watermark. The upper section, approximately 45% of the image, retains the photo's complete composition, subject, action, perspective, and real lighting. The lower section is a hand-drawn red-line rendering of the environment from the same framing. Convert the buildings, streets, plants, landforms, objects, transportation, or animal environment in the photo into monochromatic brick-red to ochre-red lines, as if drawn with thick markers or crayons on warm-white absorbent paper. Vary pen pressure and include broken strokes, retracing, cross-hatching, dry marks, and slight ink bleed. Preserve key contours and spatial relationships, remove irrelevant details, and leave broad, continuous areas of paper white. Select only the single subject in the original photo with the strongest narrative role as a realistic photographic anchor (it may be a person, animal, vehicle, food, or object). If the anchor is a person, use the fixed person reference ref-10a-v1 for identity only; do not borrow its clothing or scene. In the lower section, retain the anchor's real texture, color, material, pose, and natural shadows from the source photo, and let it touch the hand-drawn ground or environment as if stepping out of the red-line scene. Keep the anchor's count, identity, action, scale, and position consistent with the upper section. Do not turn the subject into a cartoon or add fantasy characters, fictional landmarks, fake text, brands, watermarks, original-image or AI-image labels. Do not apply a full-image filter or use smooth vector strokes.

查看 API 模型

探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作