Return to page


What is Optical Character Recognition?

Optical Character Recognition (OCR) is a process that extracts text from an image or document and converts the information into machine-readable data. The process identifies letters, converts them to words, then puts those words into sentences, enabling digital access to the original content.

How do OCR Tools Work?

OCR tools scan physical documents and convert them into black-and-white versions called bi-level document images. These scanned images are analyzed for light and dark areas. The light areas are categorized as background and the dark areas are categorized as image characters or glyphs. Glyphs are identified using two algorithms: pattern recognition and feature detection.

Pattern recognition isolates the character image, or glyph to compare examples of other stored glyphs in different formats and fonts.

Feature detection is a set of rules created for each letter or number to identify characters. Features include angled lines, crossed lines, or curves present in the character. These features are used to find the best match among stored glyphs. 

After pattern recognition and feature detection, the results are cross-referenced for accuracy using an internal dictionary.

OCR Use Cases


Financial services

The financial services industry leverages OCR to process and verify paperwork for loan documents, deposit checks, and process other financial transactions. OCR allows real-time verification of deposits via check. The OCR system creates the data that machine learning algorithms analyze to detect suspicious transactions.




The healthcare industry uses OCR to grant patients and doctors with digital access to health records, including X-rays, treatment, test results, and insurance payments. OCR enables these documents to be scanned, processed, and stored across healthcare databases. This reduces manual labor and streamlines workflow across hospitals while keeping records up to date.


Benefits of OCR

OCR automates the documentation process and improves accuracy. Time and resources are saved by removing human error from the manual documentation process.


Other benefits include:

Higher productivity

Increased accuracy

Superior data security

Text-searchable documents

Improved customer service

Easily editable documents


H2O Document AI + Optical Character Recognition

H2O Document AI, like OCR, automatically makes highly accurate AI models that extract text from an image and refines its information. H2O Document AI implements a combination of Intelligent Character Recognition (ICR) and Natural Language Processing (NLP) to leverage learning algorithms for generalizable character and word recognition that produce highly accurate and rapidly produced models.

Learn More

OCR Resources

H2O AI Document AI

Make with H2O AI Recap: Getting Started with H2O Document AI

Mission Impossible: Improving Patient Care Through Automated Document Processing