Optical Character Recognition / OCR
Technology that converts printed or handwriting text into editable digital text.
Optical Character Recognition (OCR) is a technology that converts printed or handwriting text in scanned documents, photos or screenshots into digital text, which can process and edit the computer. Traditional OCR systems recognize that when using rules-based methods that match the shape of the characters with predefined molds, today’s modern OCR systems are taking into account the context of the text by combining the evacuated nerve networks (CNN) and repetitive nerve networks or transformer architectures, which is thus achieved high accuracy in demanding texts such as corrupt, oblique or handwriting.
OCR’s importance comes from building a bridge between paper-based world and digital world: transferring invoices to accounting systems automatically, scanning ID documents, digitalization of old books in libraries, and daily use scenarios such as card or plug scan in mobile applications are always based on OCR. Google’s Tesseract is one of the most known sources of open source OCR engines, it also has gained the ability to read and understand the text directly on the image without the need for the traditional OCR pipeline.
OCR’s importance comes from building a bridge between paper-based world and digital world: transferring invoices to accounting systems automatically, scanning ID documents, digitalization of old books in libraries, and daily use scenarios such as card or plug scan in mobile applications are always based on OCR. Google’s Tesseract is one of the most known sources of open source OCR engines, it also has gained the ability to read and understand the text directly on the image without the need for the traditional OCR pipeline.