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

敦煌风格壁画中的四方天王

Four Heavenly Kings in Dunhuang-Style Murals

一幅横向敦煌风格壁画，自左至右描绘持国、增长、广目、多闻四位天王，并分别呈现白/蓝/红/绿面色、琵琶/宝剑/蛇/伞及对应竖排题字. A reusable image recipe made with gpt-image-2.5; explore the prompt and related works in the gallery.

敦煌风格壁画中的四方天王 — AI image creation
敦煌风格壁画中的四方天王

提示词 / Prompt

中文提示词

创作一幅宽幅全景壁画式绘画，借鉴敦煌佛教艺术的视觉语言，在连贯画面中清晰排列四位彼此区分的护法天王，自左至右依次呈现。使用 identity_reference ref-10a-v1 作为四位彩绘形象统一的面部身份参考，同时为每位安排不同的传统护法服饰、姿态、色彩处理和法器；不要转成现代照片，也不要添加其他人物。自左至右分别为：东方持国天王，白色彩绘面容，手持琵琶；南方增长天王，蓝色彩绘面容，手持宝剑；西方广目天王，红色彩绘面容，手持蛇形法器；北方多闻天王，绿色彩绘面容，手持宝伞。每位形象右侧分别安排独立的毛笔竖排题字，顺序为：“东方持国天王”“南方增长天王”“西方广目天王”“北方多闻天王”。四位形象及法器完整可见、间距均衡，可配以克制的祥云与织物纹样、矿物颜料色彩、古壁画质感和细致轮廓线。保持四种颜色的形象与对应题字清楚分隔。引号中的中文作为预期题字内容，不添加其他文字、印章、标志或水印。不作无依据的分辨率或相机参数承诺。

English prompt

Create one wide panoramic mural-inspired painting drawing on the visual language of Dunhuang Buddhist art, with four distinct guardian kings arranged clearly from left to right in a single continuous composition. Use identity_reference ref-10a-v1 as the consistent facial identity reference for the four stylized painted figures, while giving each a distinct traditional guardian costume, stance, color treatment and ritual implement; do not turn the scene into a modern photo or add extra figures. From left to right: Eastern King Dhṛtarāṣṭra with a white-painted face holding a pipa lute; Southern King Virūḍhaka with a blue-painted face holding a precious sword; Western King Virūpākṣa with a red-painted face holding a serpent; Northern King Vaiśravaṇa with a green-painted face holding an umbrella. Place a separate vertical brush-calligraphy label on the right side of each corresponding figure, in order: “东方持国天王”, “南方增长天王”, “西方广目天王”, “北方多闻天王”. Keep the figures and their implements fully visible and evenly spaced, with balanced decorative cloud and textile motifs, mineral-pigment colors, aged mural texture and fine contour lines. Preserve clear separation between the four color-coded figures and their labels. Treat the quoted Chinese labels as intended lettering; do not add other text, seals, logos or watermark. No unsupported resolution or camera claims.

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