あくび、背中を反る、前屈み、後ろによりかかる【nanobanana】ポーズ用のプロンプト(呪文)集｜実写画像で紹介
ai-charenge
AI美女研究所
しゃがむ、あぐら、膝を抱える、正座【nanobanana】ポーズ用のプロンプト(呪文)集｜実写画像で紹介
ai-charenge
squatting しゃがむ
indian style あぐら
hugging own legs 膝を抱える
seiza 正座
プロンプト例
Model is in a deep squat
Model is sitting cross-legged
Model sits with her knees pulled up to her chest
Model is sitting in traditional Japanese seiza.
メモ
hugging own legs（膝を抱えて座る）指示ですが、 knees pulled up to her chestのほうが再現性が高かったです。
indian style（あぐら）も、sitting cross-leggedが具体的な指示例ですね。
deepL翻訳など使いながら試行錯誤しながら試すのが近道です。
| Create ONE photorealistic 2×2 grid image. Each cell shows the SAME female model, SAME clothing, SAME hairstyle, SAME indoor room, SAME lighting. Camera angle, distance, framing, and composition MUST remain identical across all 4 cells. Only the pose changes. Global framing for all cells: Camera facing the model straight-on at seated height (not from above or below). Distance fixed: show the model from head to knees in every cell. Center framing for all cells. Neutral indoor room, soft daylight from window, warm color tone. Bed or clean floor background, no perspective changes. Photorealistic human model (NOT anime). Outfit: Same top and bottom in all cells (you may specify the outfit if needed). No color or shape changes. ================================================== TOP LEFT – “squatting” / しゃがむ Pose: Model is in a deep squat. Heels slightly raised, balancing on toes. Knees bent fully, legs close together. Torso leaning slightly forward. Both arms placed forward: elbows resting on knees OR hands together near her face, matching the reference illustration. Head slightly tilted, looking toward the camera. Same orientation as the reference “squatting” image (do NOT mirror or flip). Framing: Same camera distance and angle as other cells. ================================================== TOP RIGHT – “indian style” / あぐら Pose: Model is sitting cross-legged (“あぐら”), legs crossed comfortably. Hands resting naturally on her ankles or in front of her. Spine relaxed but upright. Body directly facing the camera, not angled. Same orientation as the reference illustration. Framing: Same camera distance and angle across all cells. ================================================== BOTTOM LEFT – “hugging own legs” / 膝を抱える Pose: Model sits with her knees pulled up to her chest. Arms wrapped around both knees. Slight forward curl of the spine. Head tilted slightly, eyes facing camera. Feet flat on the floor/bed. MUST replicate the reference pose closely (arm placement and leg position). Framing: Same camera distance and angle. ================================================== BOTTOM RIGHT – “seiza” / 正座 Pose: Model is sitting in traditional Japanese seiza. Knees together, legs folded directly under her. Feet behind her, tops of feet touching the floor. Hands placed naturally on thighs. Torso upright and centered. Same orientation as the reference illustration. Framing: Same camera distance and angle for consistency. ================================================== Strict global constraints: DO NOT zoom in or out. Model size must match perfectly across all 4 cells. DO NOT rotate, tilt, or mirror the camera. DO NOT change lighting, perspective, or background. This must look like a consistent “pose catalogue” sheet. Photorealistic only; no anime style. |
ABOUT ME
