OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image
原照转绘：高白留白彩铅水彩时尚肖像
Original Photo to Art: High-Key Colored Pencil Watercolor Fashion Portrait
Transform the uploaded portrait into a bright high-end semi-realistic fashion sketch in colored pencil and light watercolor, retaining the subject's identity, styling, vivid clothing colors and recognizable details on a predominantly white paper ground.
提示词 / Prompt
中文提示词
以提供的照片作為人物身份和造型的唯一參考，將人物轉繪成3:4 竖版高級手繪時尚肖像，整體採用半写实＋彩铅＋水彩淡彩＋时尚速写。準確保留原人物臉型、五官比例、眼眉鼻唇、表情、視線、髮型髮色、劉海、頭部角度、身體朝向、手勢、服裝、配飾及可辨識道具。不得重新設計或美化成制式AI網紅臉，不放大眼睛、不縮小鼻子、不改變臉型或嘴型。身份綁定已配置參考ref-10a-v1。臉部完成度最高：膚色清透明亮自然，保留紙白高光和淡淡紅潤；眼睛清晰有神，帶自然高光和細緻睫毛；嘴唇保持原本形狀，沿用原照片唇色或自然珊瑚粉／豆沙粉，顏色清楚但有透明感。嚴格保留原髮型與髮色。頭髮以黑、深灰、深棕彩鉛與水彩塑造深度，少量淺色點綴，並加入大量自然瀟灑的黑灰速寫飛線；線條長短、深淺、連續性各異，沿真實生長方向延伸，部分自然越出輪廓。黑髮必須明確保持黑色，不得畫成淺灰。服裝顏色忠實沿用原照：粉仍清楚是粉、藍仍清楚是藍，其他顏色亦然。以彩鉛排線、短促麥克筆筆觸、水彩淡彩和黑灰結構線共同表現服裝，保留20%–30%的白紙、斷色和未完成區域，不使用數位平塗。完整保留原照中花朵、眼鏡、耳環、髮飾、帽子、圍巾、耳機、珠寶、包等高度可辨識細節，形狀與顏色對應原照。背景改為高白度純白繪圖紙，不描繪原照片的真實環境；85%–90%背景保持純白，只在頭髮後方、肩部及身體兩側加入少量天空藍、湖藍、薄荷綠、薰衣草紫與粉色的自由彩鉛／麥克筆刷痕。色塊面積小但清晰可見，飽和度不可過低。人物邊緣保留未完成手稿感：部分髮尾、肩部、衣物下緣和身體輪廓自然斷開、留白並融入紙面。在人物旁的自然留白處加一個極小、藍／紫／粉單色線條且不填色的手繪愛心。整體必須純白、明亮、清透、鮮活且高級，手繪筆觸明顯；臉清楚但不蒼白，黑髮有深度，原服裝顏色鮮明，裝飾筆觸醒目，黑灰速寫線有力。通透不等於褪色。禁止制式AI網紅臉、動漫／韓漫、3D／CG、油畫／厚塗、傳統炭筆素描、純水彩、兒童蠟筆、粗黑統一描邊、整體低飽和灰蒙、白霧濾鏡、整體褪色、複製原照背景、複雜場景、文字、標誌與水印。
English prompt
Use the uploaded photo as the sole reference for the person's identity and styling. Create a high-end hand-drawn fashion portrait in 3:4 竖版 format using 半写实＋彩铅＋水彩淡彩＋时尚速写. Preserve the original person's face shape, feature proportions, eyes, brows, nose, lips, expression, gaze, hairstyle and color, bangs, head angle, body orientation, gesture, clothing, accessories and recognizable props. Do not redesign or beautify into a generic AI face; do not enlarge eyes or shrink the nose or alter face/mouth shape. Bind identity to configured reference ref-10a-v1. Give the face the highest level of finish: clear, bright natural skin tones, paper-white highlights and a slight natural flush; clear expressive eyes with natural highlights and delicate lashes; preserve the original lip shape and use its source color or natural coral/rosewood pink with transparent color. Keep the original hair color true. Build hair depth with black, dark gray and dark brown pencil/watercolor plus a few light accents; add abundant lively black-gray sketch flyaway lines of varied length, value and continuity, following actual growth direction and occasionally extending beyond the silhouette. Black hair must remain visibly black, never pale gray. Preserve the source clothing colors faithfully: pink stays clearly pink, blue stays blue, and likewise for other hues. Render clothing with colored-pencil crosshatching, short marker strokes, light watercolor and black-gray structural lines, leaving 20%–30% of the image as white paper, broken color and intentionally unfinished passages; no digital flat fills. Retain all highly recognizable source details—flowers, glasses, earrings, hair ornaments, hats, scarves, headphones, jewelry, bags and similar items—with matching shapes and colors. Replace the original environment with high-whiteness pure white drawing paper; do not depict the source photo's real background. Leave 85%–90% of the background pure white, adding only a few clearly visible free pencil/marker strokes in sky blue, lake blue, mint green, lavender and pink behind the hair, shoulders and sides of the body; keep the colored areas small but saturated enough to read. Preserve an unfinished-manuscript edge: selected hair ends, shoulders, lower garment and body contours naturally break, leave gaps and dissolve into paper. Add one tiny unfilled blue, purple or pink single-line hand-drawn heart in a natural blank area beside the person. Overall: pure white, bright, clear, vivid and premium, with obvious hand-drawn marks. Keep the face clear but not pale, black hair deep, original clothing colors vivid, decorative marks visible and black-gray sketch lines strong. Transparency is not fading. Prohibit generic influencer face, anime/manga, 3D/CG, oil or impasto, traditional charcoal sketch, pure watercolor, children's crayons, thick uniform black outlines, globally desaturated grayness, fog filters, faded treatment, copying the original background, complex scenery, text, logos and watermarks.
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
相关作品
- 医院公共洗手间流浪狗低机位四格视角
9:16四格写实场景，以流浪狗进入医院公共洗手间后的低机位第一人称观察视角呈现；狗不入镜。画面聚焦卫生间设施、清洁用品和一位穿实用工作服的成年护士，保持公共空间隐私与非性化表达。
- 成年女忍者透明冰蓝电子宠物机第一人称插画
3:4高调梦幻日系动画插画：成年女忍者以第一人称近景举起透明冰蓝蛋形电子宠物机，红框像素屏显示同一角色；金属钥匙圈挂满忍具与玻璃、红绳等小饰物，前后景虚化营造强烈纵深。人物使用固定参考身份，姿态克制、非情色。
- 双场景实景照片与钢笔水彩配对拼贴
竖版4:5暖白纹理纸拼贴：上半部并列户外咖啡馆与海滩两张写实照片，下半部并列对应场景的钢笔水彩插画；每一栏上下配对，分别保留英文短句。两栏人物统一为固定成年女性参考角色。
- 成年女性黄金时刻都市高端时尚杂志封面
竖版4:5高端时尚杂志封面：明确成年女性身着宽松西装、短款上衣和利落靴子，站在黄金时段现代城市街头；电影感暖光、城市散景、轻微动态模糊与极简英文封面排版。
- 现代高层写字楼玻璃中庭仰视建筑摄影
明亮的现代高层办公中庭，透明玻璃楼板与天花、白色吊顶和窗外城市天际线构成丰富纵深；从下层低角度仰拍，多位成年职场女性在远处办公、交谈或行走，保持隐私安全且不聚焦人物。
- 成年女性韩国都市路口V手势休闲街拍
4:5都市街头时尚抓拍：明确成年女性以V手势微笑，穿浅灰红心图案卫衣和白色百褶短裙，身处有蓝色巴士、车辆和黄盲道的现代韩国城市路口；阴天漫射光与自然手机摄影质感。