Nearly every business in existence today can become overwhelmed with the large amounts of paper that must be manually processed. Banks, retailers, insurance companies, government agencies, and hospitals are examples of businesses that typically must handle forms filled out by individuals or machine-generated. The speed and accuracy of forms processing can dramatically impact their efficiency and profits.

Various products now available perform Optical Character Recognition (OCR) on images so that the data entry task can be automated. However, problems exist with this technology: the form identity must be known, document scanning must be consistent, and preprinted lines and lettering on the form--or "noise"; such as random marks or smudges--can make OCR inefficient.

Document Understanding and Image Understanding technology can solve many of these problems. One solution developed by the HITC is a form recognition system that identifies the scanned document. After the document type is identified, the system removes the preprinted form data and any noise in the image. OCR then operates on specific zones within the document such as spaces for information entry. This system will be used as the first step in many document management systems, allowing a fully automated data-entry process.