人物

#使用Leonardo AI解讀圖片整體結構，在以參考人物照片，輔助生成。

#訓練完之後，提供一張參考照片，簡單的題詞：Front view photo, pure white background

Nano Banana 測試  (#需準備一張人物照)

角色： 你是一位專業的創意總監。你的目標是根據附帶的照片生成一組九宮格共計 9 張9:16尺寸具有高度一致性的影像。
第一步：角色分析
鎖定參考影像中角色的面部特徵、髮型和服裝。
第二步：環境設定
我會提供一次 [ENVIRONMENT]。你必須固定此設定去構思這個ENVIRONMENT應該有的知名場景，並將其套用於所有9個鏡頭。
第三步：9 個鏡頭序列（逐一進行）
ECU (極特寫)： 臉部的極特寫，瞳孔中帶有環境的倒影。
CU (特寫)： 肖像特寫，帶有電影般的背景虛化（景深）。
MCU (中特寫)： 半身（腰部以上），在環境中呈現自然的姿勢。
MCU (中特寫)： 半身（胸部以上），在環境中呈現自然的姿勢。
Full Body Action (全身動作)： 充滿動態感或動作的全身鏡頭。
High Angle (高角度/俯拍)： 由上往下俯瞰角色的剪影。
Low Angle (低角度/仰拍)： 由下往上仰拍，展現強大的氣場與站姿。
Long Shot (遠景)： 角色在廣闊環境中作為一個微小的身影。
Candid (抓拍)： 捕捉動作瞬間的自然反應。

操作規則：
在我提供 [ENVIRONMENT] 後，你直接用奈米香蕉PRO生成4K品質的照片，[ENVIRONMENT] =新北市新莊區天主教輔仁大學*

Nano Banana 練習 (#需準備一張人物照)

請根據此張照片，人物維持一致不變與服裝不變，請生成六個不同角度的照片。 

Put the clothes from image2 on the girl in image 1, and do not change the person's propotions. fully body. wearing black high heel. (Nano Banna)

挑一張，到GPT輸入：
我想要生成這個模特兒的電傷拍攝圖，人物跟衣服都不要改變，請幫我生成個角度的prompt提示詞，並請全部轉成英文，每個都可以讓我一鍵複製的code，加強臉部及衣服不能更動的prompt、要精簡。      (#生成可複製的code)  

回到Higgsfield一樣先上傳一張模特參考照，生成多張照片。  以此獲得不同方向的模特兒，也可請GPT生成帶有場景的code。

不同角度拍攝提示詞 Prompt

[Front — full body]

img2img from [REFERENCE_IMAGE]; same person and outfit only—do not change face, skin tone, hair, makeup, dress, white trims/bows, or heels; studio light-gray seamless, soft even light; camera at waist height, 50mm; neutral pose, arms relaxed; centered full body, natural stance, clean floor shadow.

[Front — mid shot]

img2img from [REFERENCE_IMAGE]; identity and clothing locked—no changes to face or dress details; studio light-gray, soft even light; camera at chest height, 50–70mm; crop from head to mid-thigh, centered, neutral expression.

[45° left — full body]

img2img from [REFERENCE_IMAGE]; keep exact face and outfit unchanged; model turns ~45° to camera (left shoulder slightly forward), weight on back leg; studio light-gray, soft even light; waist-height camera, 50mm; full body, centered.

[45° right — full body]

img2img from [REFERENCE_IMAGE]; do not modify face, hair, makeup, dress, bows, or heels; model turns ~45° to camera (right shoulder slightly forward), relaxed arms; studio light-gray, soft even light; waist-height camera, 50mm; full body.

[Left profile — side view]

img2img from [REFERENCE_IMAGE]; lock identity and outfit exactly; pure left profile, shoulders aligned; studio light-gray, soft even light; camera at waist height, 50–70mm; full body, centered, natural stance.

[Right profile — side view]

img2img from [REFERENCE_IMAGE]; preserve face and clothing exactly; pure right profile, shoulders aligned; studio light-gray, soft even light; waist-height camera, 50–70mm; full body, centered.

[Back — full body]

img2img from [REFERENCE_IMAGE]; same person/outfit—no changes to hair length, dress fit, or heels; model faces backdrop; studio light-gray, soft even light; waist-height camera, 50mm; full body, centered, arms relaxed.

[Back 45° — over-shoulder]

img2img from [REFERENCE_IMAGE]; identity and clothing fixed; model back 45° with slight over-shoulder glance; studio light-gray, soft even light; waist-height camera, 50mm; full body, centered.

[Upper body — neckline & bows close-up]

img2img from [REFERENCE_IMAGE]; DO NOT change face or dress details; tight crop head to waist highlighting V-neck and two white bows; studio light-gray, soft even light; camera at chest height, 70–85mm; neutral expression.

[Detail — sleeve & trim]

img2img from [REFERENCE_IMAGE]; clothing must remain identical (fabric, color, piping); macro crop of left sleeve and white trim; studio soft light, minimal shadow; 85–100mm; crisp texture, no retouching that alters design.

[Walking — natural motion]

img2img from [REFERENCE_IMAGE]; identity/outfit locked; model takes a small step forward, subtle dress movement; studio light-gray, soft even light; waist-height camera, 50mm; full body, centered, clean floor shadow.

[Hands-on-waist — shape view]

img2img from [REFERENCE_IMAGE]; keep face and outfit exactly the same; model places hands lightly on waist to show silhouette; studio light-gray, soft even light; waist-height camera, 50mm; full body, centered.

更具有動態感的場境 Prompt

[Street crosswalk — mid-step]

img2img from [REFERENCE_IMAGE]; lock identity and outfit—do not change face, hair, dress, bows, or heels; urban crosswalk, soft overcast; model takes a natural step forward, arms swinging slightly; camera waist-height, 50mm; subject sharp, light background motion blur.

[Wind gust — subtle twirl]

img2img from [REFERENCE_IMAGE]; face and clothing must remain identical; breezy sidewalk, clean background; gentle wind lifts hair and hem slightly (no design change); small pivot turn, relaxed smile; 70mm, shutter 1/500 to freeze motion; full body, vertical framing.

[Storefront walk-by]

img2img from [REFERENCE_IMAGE]; identity/outfit locked; model walks past glass storefront with faint reflections; steady stride, eyes forward; camera tracking left-to-right, 50mm; subject sharp, background streaked with mild motion blur; even daylight.

[Step-down on stairs]

img2img from [REFERENCE_IMAGE]; do not alter face or dress details; exterior stone steps; model stepping down one step, weight shifting, arms balanced; 70mm, soft directional light; full body, clean shadow, minimal crowd.

[Over-shoulder glance while walking]

img2img from [REFERENCE_IMAGE]; keep exact face and outfit; city sidewalk, shallow depth of field; model mid-walk turns head slightly over shoulder, hair in motion; camera at waist height, 85mm; crisp subject, creamy bokeh; neutral color grade.

[REFERENCE_IMAGE]=上傳的人物照片檔案

角色核心描述（Identity Anchor）
same female character, consistent face identity, East Asian female model, age 23–26, oval face, soft jawline, natural cheekbone, almond-shaped eyes, calm confident gaze, straight nose, balanced facial proportions, natural makeup, clear skin texture, dark brown eyes, medium-length black hair, slightly wavy, center-parted hair, healthy hair shine, slim fashion model body proportion, realistic human anatomy, high realism, natural lighting, editorial fashion photography style, no stylization, no anime, no illustration, photorealistic, DSLR quality, shallow depth of field

01｜正面標準肖像（基準臉）
front-facing portrait, neutral expression, eye-level camera, head and shoulders shot, studio soft light, clean background, fashion brand model test shot