Photogpt
Use Image 1 as the strict composition and scene reference. Recreate the exact same environment and framing: Inside a subway/metro train First-person perspective (viewer sitting) One hand holding a spiral sketchbook in the foreground The real person sitting in front on the seat Same camera angle, depth, framing, and positioning Do NOT change perspective, framing, or environment layout. Use Image 2 as the character identity reference. The real person sitting in front must match the face, hairstyle, and identity from Image 2 with high accuracy. 🎯 Dual-Layer Concept (CRITICAL) The image must contain TWO versions of the same person: 1. Real Person (Background) Sitting naturally on the subway seat Looking at phone (or relaxed neutral pose) Wearing casual modern outfit Face must match Image 2 exactly 2. Sketch Version (Foreground — in notebook) Pencil/ink drawing of the SAME person Clean hand-drawn portrait style Black-and-white sketch with selective color only on clothing Clothing color in sketch must match the real person’s outfit ✍️ Sketch Style Rules (VERY IMPORTANT) Pencil sketch / ink illustration style Fine linework, soft shading Minimal but clean detailing No photorealism inside the sketch Slight artistic enhancement but still recognizable 👁️ Perspective & Alignment The sketch must align with the real person (same angle and pose) Foreground sketchbook slightly tilted naturally Hand holding sketchbook visible and realistic Depth of field: sketch sharp, background slightly softer 🎥 Environment & Lighting Realistic subway interior (metal poles, seats, doors, overhead handles) Soft indoor lighting Natural shadows and reflections Slight cinematic tone but still realistic 🎨 Rendering Style Ultra-realistic photo for environment and person Hand-drawn illustration for sketch only High detail, sharp focus 8K quality Ensure the final image feels like a real candid moment where someone sketched the person sitting in front of them, with perfect identity match and perspective alignment. ❌ NEGATIVE PROMPT (VERY IMPORTANT) wrong perspective, front camera view, distorted face, identity mismatch, sketch looks photorealistic, messy drawing, extra people focus, incorrect pose, blurry, low detail, unrealistic hand, warped notebook, misaligned sketch, cartoon style

PhotoGPTのAI画像

Gemini Nano Banana 2.1
プロンプト

Use Image 1 as the strict composition and scene reference. Recreate the exact same environment and framing: Inside a subway/metro train First-person perspective (viewer sitting) One hand holding a spiral sketchbook in the foreground The real person sitting in front on the seat Same camera angle, depth, framing, and positioning Do NOT change perspective, framing, or environment layout. Use Image 2 as the character…

サイズ1856px x 2304px
参照画像
元画像 1
元画像 2

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