What Is OCR?
Optical Character Recognition (OCR) is a form of technology that identifies the characters – like numbers and letters — included in an image. Sometimes known as Text Recognition, OCR also recognizes patterns and classifies information for Artificial Intelligence (AI) to use.
For businesses, OCR can significantly improve productivity – especially for departments that handle a large number of printed documents. Once pages are processed, the text they contain can be much more readily edited, searched, indexed, and retrieved.
How Optical Character Recognition Works
To go from image to completion, OCR goes through a series of steps:
First, a paper document is read by a scanner and converted to binary data. The computer analyzes the scanned image’s light and dark areas, classifying the light areas as background and the dark ones as handwriting. This step is called Image Acquisition.
Next comes pre-processing — converting color or grayscale to binary. During segmentation, feature extraction, and classification – the next three steps — OCR finds the letters, numbers, and symbols that are inside the dark areas, using a variety of techniques — most of which examine one character or word at a time.
Then there’s pattern recognition: identifying text across multiple fonts and formats to compare characters from the scan. Assigned rules help OCR identify a single letter’s different forms: typed vs handwritten, for example, or upper vs lower case.
At the end of the OCR process, the characters are converted to ASCII or similar codes. ASCII (American Standard for Information Exchange) is the most common format for computerized text files. In this system, every letter is represented by a different 7-bit binary number.
How Optical Character Recognition Can Be Used
1. Data Processing
One popular OCR use is data entry. Companies rely on the software to convert hard copies of legal and other business documents into PDF files so employees can edit, format and search content just as they would a word processor document.
2. Data Classification
OCR can be used for data classification, which allows post offices to use the tech to sort letters and banks to electronically deposit checks.
OCR can also be used to add certified legal documents to a database, to index printouts for search engines, and to convert documents into text for the visually impaired. OCR also plays a role in translation apps, online text databases like Google Books, and security cameras that recognize license plates.
Expanding OCR With AI
Last year, an AI-based OCR for broken or partial characters debuted at the 2019 Japanese Culture and AI Symposium. This technology is especially helpful for reading kuzushiji, a form of historic, Japanese cursive that’s so different from how people write today that few can read it accurately.
OCR-driven apps are also on the rise. Working through a smartphone’s camera, these applications scan business cards and other documents, automatically loading data straight to the cloud, where people can later access it by PC or phone.