Home / Computers / OCR

OCR

Vangie Beal
Last Updated June 23, 2021 7:06 am

OCR, short for optical character recognition, refers to the technology used to convert printed, written, or typed characters into a digital format. The process allows text to be read by a computer which makes the characters able to be edited and searched. This technology has not only revolutionized data entry, but is also a helpful tool for the blind and visually impaired.

How does OCR work?

After using an optical scanner to scan a document and breaking it down into characters, OCR technology has two different approaches to converting characters to a digital format pattern recognition and feature detection.

Pattern recognition, also known as matrix matching and pattern matching, works by comparing a character that it scans to a backlog of character images that the program has on file. This only works when characters are an exact pixel match to the characters that the OCR uses as a reference. In the 1960s, the OCR-A font was created to be printed on checks in order to standardize the font for bank s OCR programs. Eventually, pattern recognition expanded to recognizing multiple common fonts like Arial and Times New Roman, but the program was still restricted to only converting characters that were in a select few fonts.

Feature detection, also referred to as feature extraction, allows OCR to read characters in almost any font. Instead of looking for a character s exact replica, feature detection works by distinguishing specific features that a character will have no matter the font. For example, if the program scans a character that has two slanting lines that meet at the top to make a point, it can tell that the letter is a capital A. Feature detection expands the uses and capabilities of OCR programs, and is the most common type of OCR used today.

Although OCR is limited to converting printed text, intelligent character recognition (ICR) can turn handwriting into text that can be read, searched, or edited by a computer. Although it is in the same family as OCR, ICR is a much more difficult process since handwriting is more widely varied than fonts.

Uses for OCR

  • Data entry
  • Assist the blind and visually impaired
  • Digitize books
  • Mail sorting
  • Make PDFs, books, and physical documents searchable
  • License plate recognition
  • Passport recognition