Photogpt
NANOBANANA PROMPT — 20 HAIRSTYLE TRY-ON (UNIVERSAL, IDENTITY-LOCKED) TASK Generate 20 different realistic hairstyles from ONE input portrait image. This prompt must work for any person (male or female), any ethnicity, any face shape. INPUT - One clear human portrait photo (front-facing or slight angle) OUTPUT - One single 5×4 grid image (20 variations total) - Same person in all images - Only hairstyle changes ──────────────────────────── IDENTITY LOCK — ABSOLUTE (CRITICAL) ──────────────────────────── Preserve WITHOUT EXCEPTION: - Face identity (facial structure, eyes, nose, lips) - Age - Skin tone and texture - Ethnicity - Facial proportions - Head size and orientation - Expression: same natural expression with a SLIGHT, relaxed smile - Eye direction and gaze - Lighting consistency on the face NO face beautification. NO identity drift. NO facial distortion. ──────────────────────────── HAIRSTYLE VARIATION RULES ──────────────────────────── - ONLY the hairstyle changes across the 20 images - Hair must be physically realistic and naturally attached to the scalp - No wig-like appearance - No floating hair - No exaggerated volume Variation scope: - Length (short / medium / long) - Cut and silhouette - Bangs / fringe / layers - Volume and styling direction DO NOT change: - Hair color (lock original hair color exactly) - Facial hair presence (if any) - Head shape ──────────────────────────── GENDER & ETHNICITY RESPONSIVENESS ──────────────────────────── - Automatically detect the subject’s gender - Generate hairstyles that are socially typical and realistic for that gender - Respect natural hair texture associated with ethnicity (e.g. straight, wavy, curly, coily) - No culturally inappropriate or unrealistic styles This must feel like a real hairstyle try-on for THIS person. ──────────────────────────── HAIRSTYLE DISTRIBUTION (20 TOTAL) ──────────────────────────── Balanced mix: - Short styles - Medium-length styles - Long styles - Clean / natural styles - Lightly styled / polished styles Each hairstyle must be clearly distinct. ──────────────────────────── CLOTHING & BODY ──────────────────────────── - Replace original outfit with a neutral studio top (plain white or light gray) - No accessories - No jewelry - No neckline variation ──────────────────────────── BACKGROUND & LIGHTING ──────────────────────────── - Pure studio background - Neutral light gray or white - Same background across all 20 images - Same lighting direction and softness in all frames - No shadows inconsistency ──────────────────────────── LAYOUT & COMPOSITION ──────────────────────────── - Final output must be a clean 5×4 grid - Even spacing between portraits - Identical framing and crop for all images - Head centered in each cell - Consistent scale and camera distance ──────────────────────────── RENDERING STYLE ──────────────────────────── - Photorealistic - Natural skin texture - Natural hair fiber detail - Looks like a professional hairstyle try-on tool - No AI-art look ──────────────────────────── FAILURE PREVENTION (STRICT) ──────────────────────────── FORBIDDEN: - Face swapping - Face reshaping - Skin smoothing - Lighting changes between frames - Makeup changes - Expression changes - Unrealistic wigs - Gender mismatch - Ethnicity mismatch ──────────────────────────── FINAL GOAL ──────────────────────────── A clean, professional 5×4 hairstyle comparison grid that makes the viewer feel: “This is exactly me — just with different real hairstyles.”

PhotoGPTのAI画像

Gemini Nano Banana Pro
プロンプト

NANOBANANA PROMPT — 20 HAIRSTYLE TRY-ON (UNIVERSAL, IDENTITY-LOCKED) TASK Generate 20 different realistic hairstyles from ONE input portrait image. This prompt must work for any person (male or female), any ethnicity, any face shape. INPUT - One clear human portrait photo (front-facing or slight angle) OUTPUT - One single 5×4 grid image (20 variations total) - Same person in all images - Only hairstyle…

サイズ1856px x 2304px
参照画像
元画像 1

創作を始める

数千人のアクティブユーザーが集まるDiscordコミュニティに参加して、最新のプロンプトアイデアやコミュニティ作品をチェックしましょう

画像を生成Discordに参加
Google Veo 3.1Seedance 2

関連作品

コミュニティの似たAI生成作品を探索できます

このスタイルを試す