English
Go to the download page

Preview: this language version was translated with machine assistance and has not been reviewed by a native speaker. Search engines are asked not to index it.

OCR (optical character recognition)

OCR, optical character recognition, reads the images of text in scans and photos as text a computer can process, and gives the position of each character on the page.

Last updated:

Role in document processing

  • Finding: without OCR, the text in a scan is only pixels and cannot be searched.
  • Locating: OCR gives the position of each character. In-place correction and redaction both depend on it.
  • Checking: read-back verification runs OCR again on the output file to check the text that was written.

Confidence

OCR gives each result a confidence score. PDF Text Replacer marks matches below 80% as “needs review” and highlights them with a warm background.

What affects accuracy

Scan resolution (300 DPI recommended), font size, typeface (handwriting and decorative fonts are recognized less reliably), background complexity, and the script itself. For character error rates on real scans in each language, see the language support matrix.

Related pages

Languages

Chinese, English, Thai and Japanese recognition is built in. Korean, Vietnamese, Arabic, Devanagari and others use language packs. Thai is shaped with HarfBuzz.

Learn more →

OCR language support matrix for scans

We tested the recognition error rate and replacement results for 17 scripts on real scanned documents. Simplified Chinese and English are primary languages; 8 scripts including Thai, Traditional Chinese and Japanese are usable but need checking; 8 are still experimental; Lao scans are not usable.

Learn more →

Text layer

The text in a PDF that you can select, copy and search. In a scan the visible text is an image, sometimes with invisible OCR text beneath it.

Learn more →