OpenZhenZhen AI プロンプトギャラリー · gpt-image-2.5 · image

同一人物の安定再現ガイド

Guide to Consistent Character Reproduction

「同一人物の安定再現ガイド」は gpt-image-2.5 を使用した作品です。OpenZhenZhen で元のプロンプトと生成例を確認し、創作の参考にできます。

同一人物の安定再現ガイド — AI image creation
同一人物の安定再現ガイド

プロンプト

中国語プロンプト

以用户上传的参考图作为主要可见人物的面部身份依据。在不与身份参考冲突的情况下,保留来源的场景、造型、服装、姿势、构图、光线和视觉风格。 基于附图中的人物,今后为稳定再现同一人物而制作的“人物取扱说明书”,请以横长1页的形式创建。 以附图为唯一标准,在所有面板中统一面容、年龄感、体型、发型、发色、服装、皮肤质感。禁止变成他人、过度美化、年轻化。 放入シート的内容: ・全身:正面、侧面、后方、斜前方、斜后方 ・脸部:正面、侧脸、斜45度、上方、下方 ・表情:通常、笑脸、认真、惊讶、害羞、苦脸、思考中、悲伤、平静 ・细节:眼睛、眉毛、鼻子、嘴巴、耳朵、轮廓、皮肤、手 ・头发:刘海、侧发、后发、发梢、发流向 ・头发和脸部的其他角度 ・简短特征备忘 ・头发、瞳孔、皮肤、嘴唇、服装、阴影的颜色样本 ・再生成时的固定要点 横长16:9,白色或浅灰色背景。以细薄的横线和方框整理的、干净的资料用设计来制作。请将全身和脸部置于中央,表情和其他角度置于右侧,部件细节置于下方。 图像内的文字仅限于简短的日语标签,不要加入姓名、年龄数字、性格、爱好、简介等。原图的实拍、插图等表现形式也请保持原样。

英語プロンプト

Use the uploaded reference image as the facial identity reference for the main visible person. Keep the source scene, styling, clothing, pose, composition, lighting and visual style unless they conflict with the identity reference. Based on the character in the attached image, please create a 'Character Instruction Manual' in a wide single-page format for the stable reproduction of the same character in the future. Using the attached image as the sole standard, unify the face, sense of age, body type, hairstyle, hair color, clothing, and skin texture across all panels. It is forbidden to turn into someone else, over-beautify, or make them look younger. Contents to be included in the sheet: ・Full body: Front, side, back, diagonal front, diagonal back ・Face: Front, profile, 45-degree angle, above, below ・Expressions: Normal, smiling face, serious, surprised, shy, bitter face, thinking, sad, calm ・Details: Eyes, eyebrows, nose, mouth, ears, contour, skin, hands ・Hair: Bangs, side hair, back hair, hair ends, hair flow direction ・Other angles of the hair and face ・Short characteristic memo ・Color samples for hair, pupils, skin, lips, clothing, and shadows ・Fixed points for regeneration Landscape 16:9, white or light gray background. Create with a clean document-style design organized with thin horizontal lines and boxes. Please place the full body and face in the center, expressions and other angles on the right, and part details at the bottom. Text within the image is limited to short Japanese labels; do not add names, age numbers, personality, hobbies, profiles, etc. Please also keep the expression style of the original image, such as photorealistic or illustration, as it is.

API モデルを見る

画像・動画プロンプトをもっと見る · 画像を生成 · 動画を生成 · AI キャンバスで作成