OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image
十五格头部转向与俯仰人像参考表
15-Panel Head Turn and Tilt Portrait Reference Sheet
以5列×3行的整齐矩阵展示同一人物的15张头肩肖像:五列依次为左侧面、左四分之三、正面、右四分之三、右侧面;三行依次为仰头、水平、低头。每列跨三行保持相同水平转角,只改变头部俯仰,并维持身份、发型、服装、配饰与光线一致。

提示词 / Prompt
中文提示词
创作一张干净的暖象牙色人像参考表,以严格的5列×3行网格呈现恰好15张一致的头肩肖像,格间距均匀且窄。所有画格均使用已配置的ref-10a-v1保持同一人物身份。每格的面部比例、发型、服装、配饰、相机透视和灯光一致。每一列固定一种水平转头方向;每一行固定一种垂直俯仰角度。 从左到右五列依次为:(1) 清晰左侧面,呈现鼻、唇、下巴、下颌线、耳朵和颈部轮廓;(2) 左侧四分之三角度,可见双眼并呈现清楚的面部纵深;(3) 双侧面部特征均衡的正面;(4) 右侧四分之三角度,可见双眼并清楚呈现另一侧脸颊;(5) 清晰右侧面,呈现面部与颈部轮廓。 从上到下三行依次为:第一行——五个头部全部向上仰,抬起下巴,面部角度高于水平中性位;第二行——五个头部全部自然端平,不刻意仰头或低头;第三行——五个头部全部向下俯,下巴低于中性位。每一列在三行中都必须保持原定的左侧面、侧前方、正面或右侧面水平转角。对应画格之间唯一主要姿势差异是仰头、水平或低头。额鼻线、下巴、下颌线和颈部姿态都要清楚表现俯仰差异,便于对照;不要用斜向歪头代替俯仰,也不要跨行改变水平转角。 视线方向须符合各角度的解剖关系:正面和四分之三角度在合理时可看向镜头,但不要让侧面人物不自然地扭转眼睛看向观者。颈部活动与下颌至颈部轮廓保持真实自然,避免极端后仰、压缩或扭转。每行严格五张、总计三行。不得增加、遗漏、合并或重排画格;不要随机转头、同一行混合不同俯仰、改变相机透视或造成身份漂移。
English prompt
Create one clean warm-ivory reference sheet containing exactly 15 consistent head-and-shoulders portraits in a strict 5-column × 3-row grid, with uniform narrow spacing. Use configured ref-10a-v1 for the same person's identity in every panel. Keep the same facial proportions, hairstyle, clothing, accessories, camera perspective, and lighting throughout. Each column fixes one horizontal head rotation; each row fixes one vertical head pitch. Columns from left to right: (1) clear left-facing profile showing nose, lips, chin, jawline, ear, and neck silhouette; (2) left three-quarter view with both eyes visible and clear facial depth; (3) balanced frontal view; (4) right three-quarter view with both eyes visible and the opposite cheek defined; (5) clear right-facing profile showing the facial and neck silhouette. Rows from top to bottom: Row 1—every head tilted upward, chin raised and face angled above neutral. Row 2—every head upright and naturally level, with no deliberate upward or downward pitch. Row 3—every head tilted downward, chin lowered below neutral. Within each column, preserve its assigned left/profile/three-quarter/frontal/right orientation across all three rows. The only major pose change between corresponding panels is upward, neutral, or downward head pitch. Make the pitch differences visibly comparable in the forehead-to-nose line, chin, jawline, and neck posture; do not substitute diagonal roll or change horizontal rotation between rows. Keep gaze anatomically natural for each view: frontal and three-quarter eyes may look toward camera where plausible, but do not twist profile eyes unnaturally toward the viewer. Use realistic neck articulation and natural jaw-to-neck contours; avoid extreme extension, compression, or twisting. Maintain exactly five portraits per row and three rows total. Do not add, omit, merge, or rearrange panels; no random orientations, mixed pitches within a row, inconsistent camera perspective, or identity drift.
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
- 明亮CCD都市通勤生活写真
A vertical daytime CCD-style urban commuter portrait of an adult woman emerging onto a modern subway plaza in a sky-blue knit top and cream-white short shorts.
- 持银色数码相机的深发成年女性人像
A clean 3:4 vertical portrait of an adult dark-haired woman holding a compact silver camera against a softly blurred dark glass facade with warm reflected lights.
- 孔版印刷超大咖啡杯通勤插画海报
以成人通勤者在公交站等车为日常场景,将一个外带咖啡杯夸张放大,采用严格三色孔版印刷质感和9:16竖幅构图。
- 水色短款上衣与格纹裙的巷中街拍
一位成年女性在狭窄都市小巷中的克制街头时尚人像,穿水色短款上衣与格纹裙,采用3:4纵向构图。
- 红蓝涂鸦风成年女性动漫动作海报
以极端低机位表现一位成年女性动漫角色俯身持枪对准镜头的先锋动作海报,融合丙烯、街头涂鸦与高反差商业插画质感。
- 汤姆猫 · 奋力攀登 · 双色凝胶
Moss Image 原创创意图集|双色凝胶|汤姆猫。原创提示词生成。