Galerie de prompts IA OpenZhenZhen · gpt-image-2.5 · image

Portrait de face : redessin comparatif en bandes, manhua moderne

Frontal Portrait: Modern Chinese Comic Split Comparison Redraw

Portrait de face : redessin comparatif en bandes, manhua moderne — découvrez cette création avec gpt-image-2.5 sur OpenZhenZhen, son prompt original et un exemple de génération à réutiliser.

Portrait de face : redessin comparatif en bandes, manhua moderne — AI image creation
Portrait de face : redessin comparatif en bandes, manhua moderne

Prompt

Prompt chinois

将随附 ref-10a-v1 整幅 JPG 作为实际输入照片,以及唯一的人物身份与场景参考。制作一张竖版上下对比插画,采用现代中国漫画与当代国产二维动画角色设定风格。上下区域约各占一半,交界处有清楚的水平边界。上半部完整保留真实照片及其中女性、单人数量、身份、正面姿态、视角、空间关系、深青蓝场景、原有光线、材质和可识别色彩。保持摄影效果,不插画化、不加滤镜;不得改变地点或增删关键主体。下半部将同一位女性转译为现代国漫/二维动画角色,同时保留可辨识身份、正面动作与姿态。保留她实际穿着的黑色细肩带上衣和深棕色长发;不要改成汉服、添加第二个人或虚构道具。使用克制流畅的深色细轮廓线、古典修长的眉眼、准确概括的面部结构、清晰平涂色块,以及低饱和青绿、米白和少量从原照片提取的辅助色。阴影只用一至两级干净色面,不做写实明暗。将真实头发走向及肩部/衣物轮廓风格化为少量飘带般长弧线与克制的书法式动势线,同时保持她静止的正面姿态可辨。下半部背景简化为大面积暖白纸面与青绿色块,仅加入少量取自真实青蓝背景的轮廓;不堆叠景物,不虚构地点。只加入极轻微纸面颗粒,使画面像当代中国漫画角色设定稿或国产二维动画关键帧。不要文字、标签、水印或“原图/AI图”角标。禁止水墨画、工笔重彩、敦煌壁画、古画卷轴、堆叠祥云、写实厚涂、3D、日漫大眼、欧美漫画、照片滤镜、左右并排、九宫格、改变人物或姿态、复制人物和过度仙侠特效。

Prompt anglais

Use the entire attached ref-10a-v1 JPG as the actual input photograph and sole identity and scene reference. Create one vertical top-and-bottom comparison illustration in a modern Chinese comic and contemporary domestic 2D animation character-design style. The two areas occupy about half each and meet at a clear horizontal boundary. In the upper half, preserve the complete real photograph and its woman, single-person count, identity, frontal pose, perspective, spatial relationships, deep teal-blue setting, original light, material and recognizable colors. Keep it photographic, not illustrated or filtered; do not change location or add/remove key subjects. In the lower half, reinterpret this same woman as a modern Chinese-comic/2D-animation character while retaining her recognizable identity, frontal action and pose. Keep her actual modern black thin-strap top and long dark-brown hair; do not change her into Hanfu, add another person, or invent props. Use restrained, smooth dark fine outlines, classical slender brows and eyes, accurate simplified facial structure, clear flat-painted color blocks, low-saturation cyan-green and cream-white with only a few supporting colors sampled from the photo. Use only one or two clean color-plane levels for shadows, not realistic shading. Stylize the real hair fall and shoulder/clothing contours into a few flowing streamer-like arcs and restrained calligraphic motion lines, while keeping her still frontal pose recognizable. Simplify the lower background into broad warm-white paper and cyan-green planes with only a few contours derived from the actual teal backdrop; do not stack scenery or invent a location. Add only very slight paper grain, like a contemporary Chinese comic character-design sheet or domestic 2D animation keyframe. No text, labels, watermarks, or Original Photo/AI Image tags. Prohibit ink painting, gongbi heavy color, Dunhuang murals, ancient scrolls, stacked auspicious clouds, realistic thick paint, 3D, Japanese manga big eyes, Western comics, photo filters, side-by-side layouts, nine-grid layouts, changed people or poses, duplicated people, or excessive xianxia effects.

Voir le modèle API

Découvrir plus de prompts image et vidéo · Générer des images · Générer des vidéos · Créer sur le canevas IA