mcpField
P
photo-restoration

photo-restoration

github.com/shinchvenCreative & media7253 installsUpdated 35w ago
Repository

About photo-restoration

photo-restoration is a skill for directing image generation to restore blurry, vintage, or low-quality photographs. It guides prompt construction and uses the supplied image as a reference while aiming to preserve identity, pose, and composition.

Key features

  1. 01Analyze blur, grain, focus, and lighting
  2. 02Construct a restoration prompt
  3. 03Sharpen details and adjust lighting
  4. 04Preserve identity, pose, and composition
  5. 05Use the uploaded image as a reference

Use cases

  1. 01Restore a blurry photograph
  2. 02Modernize the appearance of a vintage image
  3. 03Upscale and sharpen a low-quality picture

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/shinchven/nano-banana-skills --skill photo-restoration

After installing, type/photo-restorationin Claude Code to confirm it loaded.

Details

From the project README

Photo Restoration

Capability

Transform blurry, vintage, or low-quality images into crystal-clear, 8k resolution modern photographs while preserving the original identity, pose, and composition.

Triggers

  • "Restore this photo."
  • "Fix this blurry image."
  • "Make this vintage photo look modern."
  • "Upscale and sharpen this picture."

Instructions

  1. Analyze the User's Image:

    • Note the subject's identity, pose, and composition.
    • Identify areas needing restoration: blur, grain, focus, lighting.

  2. Construct the Image Generation Prompt:

    • Core Command: "High-fidelity restoration of this vintage photograph. Transform this blurry image into a crystal-clear, 8k resolution modern photograph."
    • Enhancements: "Sharpen all details, remove film grain, and correct focus. Enhance facial features to be lifelike and highly detailed, ensuring realistic skin texture and eye clarity. Fix lighting to be balanced and natural."
    • Preservation: "Upscale to high definition while preserving the original identity, pose, and composition."
    • Style: "Professional portrait photography, shot on a 85mm lens, f/1.8, photorealistic, cinematic lighting."
    • Constraint: "While largely improving the quality of the photo, the restored image should remain identical to the original."

  3. Generate the Image:

    • Use the image generation tool.
    • Pass the constructed prompt.
    • Pass the user's uploaded image path in the reference image parameter to use as a reference/control.

Tools / Commands

  • Image generation tool: To generate the restored image based on the input.

Examples

User: "Restore this old photo." (User attaches [image]) Action:

  1. Construct prompt using the standard restoration template.
  2. Call the image generation tool with the constructed prompt and image paths.

Comments

Markdown supported