OCR for Scanned Documents: Making Old PDFs Accessible

What is OCR?

Optical Character Recognition (OCR) is a technology that converts images of text into machine-readable text. When you scan a physical document or receive a “scanned PDF,” the file contains pictures of the text — not actual text characters. A screen reader cannot read it, you cannot search it, and you cannot change its font.

OCR software analyzes the image, identifies characters based on their shape, and outputs the corresponding text. The quality of OCR output depends on the clarity of the original scan, the quality of the OCR engine, and the complexity of the document layout.

Why scanned PDFs are problematic for accessibility

For dyslexic readers, scanned PDFs represent one of the most significant accessibility barriers in education and professional life:

• The text is an image, so font changes are impossible
• Screen readers and text-to-speech tools cannot read the content
• The font is whatever was on the original printed page (often a dense serif font like Times New Roman)
• Zoom quality degrades rapidly compared to vector text
• Copy-paste does not work, preventing the use of reading aids

Educational settings are full of scanned PDFs: photocopied worksheets distributed as PDFs, scanned library textbook chapters, historical primary sources, and handwritten notes. Students with dyslexia are disproportionately affected by this inaccessibility.

How modern OCR works

Early OCR systems used pattern matching — comparing shapes in the image to templates of known characters. Modern OCR engines use machine learning and neural networks, which dramatically improves accuracy on varied fonts, scan quality, and layouts.

The best modern OCR engines (including the technology used by DysFont) can achieve 99%+ accuracy on clean scans. Key factors in OCR quality include:

• Scan resolution: 300 DPI minimum for good accuracy; 600 DPI for small text or complex layouts
• Image contrast: Clean black text on white background yields best results
• Page orientation: Straight, correctly oriented pages (OCR can auto-correct minor skew)
• Font type: Printed fonts are recognized more accurately than handwriting
• Language: OCR engines are trained per language; specifying the correct language improves accuracy

DysFont’s OCR pipeline

DysFont integrates OCR directly into its conversion pipeline. When you upload a scanned PDF or an image file, the process is:

The output is a fully searchable, accessible PDF with real text formatted in your chosen dyslexia-friendly font. It can be read by screen readers, searched, and printed normally.

OCR accuracy: what to expect

OCR accuracy varies significantly by scan quality. Here’s a practical guide:

Best practices for preparing documents for OCR

If you have control over the scanning process, these steps will significantly improve OCR accuracy and the quality of the final accessible PDF:

• Scan at 300 DPI minimum: Most modern document scanners default to 200 DPI; change this to 300 or higher in scanner settings
• Use black and white mode for text documents: Color scanning adds no benefit for text-only documents and increases file size
• Ensure pages are flat and straight: Curved pages from book scanning significantly reduce accuracy; use a book scanner if possible
• Clean the scanner glass: Dust and fingerprints on the scanner surface appear as artifacts in the scan
• Remove staples and paperclips before scanning
• For smartphone scanning: Use apps like Microsoft Office Lens or Adobe Scan which apply automatic deskewing and contrast optimization

Use cases: who benefits most from OCR + dyslexia fonts

Students and university libraries

Educational institutions frequently provide course materials as scanned PDFs — particularly older textbooks, journal articles, and historical sources. Students with dyslexia can use DysFont to convert these to accessible formats without requiring the institution to provide special accommodations for each document.

Legal and administrative documents

Contracts, legal briefs, and government documents are often distributed as scanned PDFs. Converting these to accessible formats allows dyslexic professionals to read and review them independently.

Personal archives and family history

Old letters, newspaper clippings, and family documents that have been digitized can be made readable through OCR conversion.

Historical and archival research

Academic researchers working with digitized historical texts can convert these to dyslexia-friendly fonts for more comfortable extended reading sessions.

OCR and accessibility compliance

Providing accessible versions of documents is increasingly a legal requirement. In the EU, the Web Accessibility Directive requires public sector bodies to provide accessible digital content. In the US, Section 508 and ADA Title III require that educational materials be accessible to students with disabilities.

A scanned PDF that cannot be read by screen readers or reformatted for dyslexic students does not meet these requirements. OCR conversion to accessible PDF is one practical way to bring legacy documents into compliance. See our guide on accessibility compliance for more details on legal requirements.

OCR + accessibility: the complete pipeline

Most OCR tools stop at text extraction. DysFont continues where they stop — applying the full accessibility pipeline on top of the extracted text. When you upload a scanned document, four things happen automatically:

1. OCR extraction: Image text is converted to machine-readable characters (98–99% accuracy for clean scans)
2. Spacing optimization: Letter spacing is set to 35% of average letter width — the BDA-recommended standard for dyslexia accessibility
3. Font substitution (optional): The extracted text is rendered in your chosen accessibility-friendly font
4. Color overlay (optional): Background color applied to reduce visual stress (cream, blue soft, green soft, dark mode)

The result: a 50-year-old scanned textbook becomes a fully searchable, screen-reader-compatible, dyslexia-optimized PDF in seconds. Manual remediation of the same document would take hours. DysFont does it automatically.

Why this pipeline matters for schools and institutions

Educational institutions often have extensive archives of scanned materials — photocopied worksheets from the 1990s, digitized library books, historical primary sources. These are completely inaccessible to dyslexic students in their raw form. The DysFont pipeline converts them to accessible formats that comply with BITV 2.0, Legge 170/2010, UK Equality Act, and RGAA requirements.

A library of 500 scanned PDFs that took weeks to produce can be made fully accessible in an afternoon. No specialist software, no expert knowledge required — just upload and convert.

Frequently asked questions