Photogpt
[ROLE]
You are a high-end surreal fashion art director specializing in dense multi-instance compositions and geometric studio illusions.

[IDENTITY LOCK]
Use the uploaded full-body reference image as the absolute identity source.

Strictly preserve:
- facial structure
- facial proportions
- skin tone
- body proportions
- silhouette
- clothing color and material

All instances must be the exact same person.
No identity variation allowed.

---

[COLOR PALETTE LOCK — CRITICAL]

Extract dominant palette from identity reference:

- PRIMARY (dominant clothing color)
- SECONDARY (support tones)
- ACCENT (minor highlights)

Rules:

- environment uses PRIMARY or SECONDARY (slightly desaturated or shifted for contrast)
- frames harmonize with palette (warm/cool matched metallic or coated tone)
- no unrelated colors
- full scene must feel like one unified palette system

---

[ENVIRONMENT LOCK]

- geometric studio corner (two walls + floor visible)
- sharp edges, clean intersection
- flat, textureless surfaces
- monochrome or dual-tone palette derived from subject
- no props except frames

---

[FRAME SYSTEM]

- multiple frames embedded into walls and floor
- frames act as portals
- varied sizes (small / medium / large)
- varied orientations (tilted, rotated, upside-down)
- depth layering: foreground frames + background frames

---

[COMPOSITION SYSTEM — DENSITY CONTROL]

- 6–8 instances of the same person
- chaotic scattered but **intentionally layered**
- enforce spatial hierarchy:

  • 1 primary (hero) instance — largest, clearest  
  • 2–3 secondary instances — medium scale  
  • 3–4 tertiary instances — smaller, partial or distant  

- distribute across:
  - walls
  - floor
  - mid-air

- maintain **clear separation** (no merging bodies)
- avoid uniform spacing (organic chaos)

---

[INTERACTION RULE]

Each instance must:
- emerge from or enter through frames or flat surfaces
- include variety:
  - full body
  - half body
  - limbs only
  - upside-down or rotated entries

- no duplicated pose

---

[CAMERA]

- wide-angle lens
- slight distortion for spatial exaggeration
- dynamic composition (non-centered framing)
- depth emphasized (foreground / midground / background)

---

[LIGHTING]

- studio lighting
- strong directional light
- sharp, clean shadows
- shadows must reinforce geometry and depth layering
- lighting temperature aligned with palette

---

[STYLE]

- hyper-realistic human rendering
- surreal but physically convincing
- playful / humorous tone
- high-end commercial fashion photography quality

---

[PRIORITY]

composition hierarchy > color harmony > environment accuracy > identity consistency > realism

---

[NEGATIVE]

- identity inconsistency
- face variation across instances
- merged bodies
- overlapping faces
- distorted anatomy
- duplicated poses
- flat composition (no depth)
- weak hierarchy
- random spacing grid
- clutter
- color mismatch
- texture noise
- low detail

PhotoGPT 的AI 图片

提示词

[ROLE] You are a high-end surreal fashion art director specializing in dense multi-instance compositions and geometric studio illusions. [IDENTITY LOCK] Use the uploaded full-body reference image as the absolute identity source. Strictly preserve: - facial structure - facial proportions - skin tone - body proportions - silhouette - clothing color and material All instances must be the exact same person. No…

尺寸1856px x 2304px
参考图片
源图片 1
源图片 2

开始探索

加入拥有数千名活跃用户的 Discord 社区,了解最新提示词灵感并探索社区作品