探索这项精选AI 图片 | PhotoGPT AI 社区画廊

PhotoGPT 的AI 图片

Grid prompt

提示词

{ "prompt": "Use the uploaded reference image ONLY for the woman’s exact identity, facial structure, eye color, lips, freckles, skin texture, hair appearance, and overall facial consistency. DO NOT copy the pose, framing, posture, camera angle, or expressions from the reference image. Instead, recreate the exact composition style, framing, posture positioning, and facial alignment of the attached 3x3 collage…

尺寸1216px x 1536px

参考图片

${ "prompt": "Use the uploaded reference image ONLY for the woman’s exact identity, facial structure, eye color, lips, freckles, skin texture, hair appearance, and overall facial consistency. DO NOT copy the pose, framing, posture, camera angle, or expressions from the reference image. Instead, recreate the exact composition style, framing, posture positioning, and facial alignment of the attached 3x3 collage example image. Every selfie frame must have the woman positioned centered inside each grid box with matching head placement, crop level, shoulder framing, camera distance, and facial orientation exactly like the example collage. Create a realistic 3x3 selfie contact-sheet collage featuring the same woman in nine different close-up expressions while preserving identical lighting and environment across all frames. Wet messy auburn hair, natural glossy skin, minimal makeup, white tank top straps visible, direct flash photography, candid bathroom/bedroom selfie aesthetic, raw unfiltered beauty, realistic pores and skin texture, soft shadows, authentic iPhone front-camera feel. Expressions include: angry snarl, playful tongue out, soft resting face with hands together, dreamy upward gaze with hands in hair, subtle confident smile, seductive stare with lip bite, fish-face pout, shocked open-mouth reaction, sleepy emotional face resting on hand. The posture, camera perspective, face position, framing symmetry, and centered composition in every grid cell must closely match the attached collage reference while ONLY the identity comes from the uploaded person reference image. Tight framing, intimate composition, realistic asymmetry, natural imperfections, flash reflection in eyes, Tumblr/Instagram dump aesthetic, photorealistic, ultra detailed, cinematic skin texture, realistic wet hair strands, authentic candid energy, no glamour retouching, no AI-looking skin, no cartoon style, no beauty filter look. Maintain identical environment and styling across all 9 frames like a real spontaneous selfie collage shot in one session.", "style": "photorealistic", "aspect_ratio": "1:1", "quality": "high", "lighting": { "type": "direct_flash", "mood": "raw candid nighttime selfie", "shadows": "soft realistic" }, "camera": { "lens": "iPhone front camera", "framing": "tight centered close-up selfie", "composition": "3x3 collage grid matching attached reference layout" }, "details": { "skin_texture": "natural pores and freckles", "hair": "wet messy auburn hair", "makeup": "minimal natural makeup", "consistency": "same person across all frames", "pose_reference": "match attached collage posture and face placement exactly", "identity_source": "only from uploaded reference person" } }$

photogpt5个月前

PhotoGPT 的AI 图片

开始探索

相关作品