WHAT IS OPTICAL CHARACTER RECOGNITION (OCR)?
OCR stands for Optical Character Recognition. OCR is technology that is used to recognize text that has been turned into an image. More often these images are scanned documents (such as a PDF) and photos. The OCR technology converts virtually any kind of images containing written text (typed, handwritten or printed) into machine-readable text data.
OCR technology gained widespread usage in the early 1990s when there was an increasing desire to digitize historic documents. In the interim, OCR technology has been vastly improved and today can deliver near perfect OCR accuracy.
How Does OCR Work?
Look around you right now, and you’ll see a wide variety of text. You’ll see printed text using regular Latin fonts, cursive fonts, novelty fonts, and perhaps even non-Latin (Chinese, Greek, Indic, etc.) fonts. You’ll see handwriting – your own, and that of others. We’re almost always surrounded by a multitude of text. Asking any technology to recognize all these different texts is asking a lot.
In essence, there are two different ways to accomplish this. The technology can either recognize characters in their entirety (called pattern recognition) or by detecting the individual lines and strokes that the characters are made from (called feature detection).
OCR programs must be taught to recognize letters that have been typed in a number of very common fonts (ones like Times, Helvetica, Courier, and so on). Luckily, most fonts share very similar characteristics. This goes for both Latin and non-Latin fonts. However, there is still no guarantee they could recognize any font you might send their way.
A more accurate tool in OCR is feature detection. Instead of recognizing, for example, the complete pattern of a W, pattern detection will detect the individual component features (angled lines, crossed lines, etc.) which comprise the character. It looks for features, instead of patterns, which allows it to recognize characters with a greater level of accuracy.
Handwriting, of course, presents its own set of problems. When an OCR program is asked to digitize handwriting, it often uses both pattern recognition and feature detection, and may even require additional information, such as the context of the written text. Is it a grocery list? Or a will? Is it in English? Or Hebrew?
How is OCR Used?
The most common use of OCR technology involves converting printed paper documents into machine-readable text documents, which can then be edited with word processors like Microsoft Word or Google Docs.
OCR technology is often used behind-the-scenes, powering many well-known systems and services in our daily life, such as data entry automation, indexing documents for search engines, automatic license plate recognition, and even assisting blind and visually impaired persons.