Saving Blurry Docs: Cleaning Scanned Noise, Stains, and Old Handwriting

Restoring Scanned Documents: Cleaning Backgrounds and Artifacts

Scanned PDF documents often contain visual defects—such as yellowed pages, dust spots, copier lines, or handwritten notes.

These artifacts make files look unprofessional, consume excess ink when printing, and reduce the accuracy of Optical Character Recognition (OCR) software.

Using image enhancement and thresholding algorithms allows you to clean up these scans, restoring clarity to old documents.

---

Three Core Algorithms for Document Image Enhancement

Cleaning a scanned PDF involves processing the embedded page images directly on your computer:

1. Adaptive Thresholding

Standard thresholding applies a single brightness limit to the entire page, which can turn unevenly lit areas (such as shadows from mobile photos) completely black.

Adaptive thresholding calculates local average brightness values for small pixel clusters, deciding whether each point represents text (black) or background (white). This removes shadows and yellowing, restoring the page background to pure white (`#FFFFFF`).

2. Despeckle Filtering

Dust on scanner glass creates small black dots on pages. Despeckle algorithms identify these isolated pixel clusters and remove them, cleaning up the page.

3. Text Contrast Enhancement

For faded text, edge-detection algorithms trace text outlines and darken the interior pixels, making characters stand out and improving OCR scan rates.

#### The Advantage of Client-Side Processing

Cloud-based cleaning tools often convert PDFs to JPEGs, apply corrections, and re-compress them, introducing compression artifacts. Processing files locally via WebAssembly updates the image pixel buffer directly, maintaining text sharpness.

Summary

Cleaning scanned documents improves text clarity, reduces printing costs, and enhances search accessibility, preserving the quality of your digital archives.