OCR stands for Optical Character Recognition. OCR is a software tool that can extract print text from some documents.
When will OCR work well?
OCR does not work on handwriting. It only works for printed or typed text, meaning text created by a typewriter, printing press, or other mechanical means. OCR will do best on consistent and clear images of modern typefaces.
Do I still need to review pages started with OCR?
Yes! OCR is imperfect. It may not work well for some or all parts of a typed page, but it can be a great starting point. If you start a page with OCR, you should read the text closely before submitting. If you are reviewing a OCR-ed page, you also still need to review.
We always want to use volunteer time effectively. When the Library of Congress digitizes a large group of printed pages, it will usually OCR them. The materials in By the People campaigns are not good candidates for applying OCR at scale, either because they are handwritten, a mixed collection of handwritten and print materials, or printed on paper or in a typeface that does not produce accurate OCR results. However, OCR can still be a useful starting point for some typed pages. Use it if it if you like it or skip it if you don’t!
Clicking "Transcribe with OCR" will remove all existing transcription text and replace it with automatically generated text. We recommend saving existing text in a separate document if you may want to revisit it.