Промт для восстановления и колоризации архивного портрета в ИИ

ARCHIVAL PHOTO RESTORATION WITH CONTROLLED RECONSTRUCTION

Task: Restore and colorize the photograph by repairing physical damage while preserving identity and original content.

HARD CONSTRAINTS:
- IDENTITY LOCK: Faces, proportions, expression, and pose must remain identical.
- NO IDENTITY CHANGE.

CONTROLLED RESTORATION:
- Reconstruct ONLY damaged areas (cracks, stains, tears).
- Keep reconstruction minimal and consistent with nearby pixels.
- Do NOT invent new objects or anatomy.

FACE RULES:
- Improve clarity and sharpness slightly.
- No beautification or structural changes.
- Preserve natural skin texture.

RESTORATION:
- Remove cracks, dust, stains, scratches, mold.
- Repair damage realistically.
- Restore contrast and tonal balance.
- Avoid overprocessing.

COLORIZATION:
- Natural, historically plausible colors.
- Soft, muted vintage palette.
- Natural skin tones.

LIMITS:
- Do not add/remove subjects.
- Do not change composition or framing.

TEXTURE:
- Preserve grain and micro-texture.
- Avoid strong smoothing.

SECONDARY SUBJECTS:
- Apply same restoration quality to all visible faces.

OUTPUT:
High-resolution, clean, naturally restored archival photo.

IMAGE DESCRIPTION:
1.  **Number of subjects**: One human subject, with a partial view of a doll.
2.  **Approximate age and gender**: A young child, approximately 3-5 years old. Gender is