Evidence-Aware Restoration: repair damage without inventing history

Image Input
Prompt result
Ready prompt
Use case: conservative archival photo restoration
Input image: the attached image, the damaged or degraded photograph
Restoration goal: remove scratches and restore contrast
Known context: identity and location are confirmed

Primary request: Restore the attached image according to remove scratches and restore contrast, using identity and location are confirmed only as supporting context.

Damage repair: Remove scratches, dust, stains, fold lines, tears, fading, chemical discoloration, noise, and scanning artifacts while preserving authentic photographic texture.

Tone and detail: Recover balanced exposure, contrast, local sharpness, and tonal separation without creating a modern overprocessed look. Preserve original grain and lens character where they remain visible.

Identity protection: Preserve the exact facial structure, expression, age, hairstyle, body shape, clothing, jewelry, and relationships between people. Reconstruct missing areas conservatively from nearby evidence and symmetry only when justified.

Text and evidence: Preserve legible writing and signage exactly. If letters are missing or unreadable, leave them unresolved rather than inventing words. Do not add people, objects, decorations, or background history unsupported by the image.

Color: Colorize only when the restoration goal explicitly requests it. Use restrained, period-plausible colors and keep ambiguous colors neutral.

Constraints: Avoid beauty retouching, face replacement, invented eyes or teeth, waxy skin, excessive sharpening, fake high-frequency detail, or altered composition. Add no watermark.

Output: Return only one carefully restored image.

Prompt details

Restore a damaged or degraded photograph conservatively, protecting identity and avoiding unsupported facial or textual invention.

Shared by
@picoprompt Official
Works best with
Image model with restoration editing support
Published
Sep 10, 2026