What is OCR?

OCR (Optical Character Recognition) refers to any technology that is used to recognize the text within an image-based item. 

Classification of Document Types

Media types, with regards to OCR, can be broken down into two primary categories—

Text-Based Media— An example of this type of media would be if a Microsoft Word Document is created, saved, and re-opened. When the document is re-opened, the text within it is 100% accurate. This is due to the document text being stored within the properties of the file. Document content is rendered at 100% due to being rendered purely from “rendering instructions.” Most OCR solutions are programmatically designed to read these rendering instructions and can therefore output the document text with 100% accuracy.

Image-Based Media—Image-based media, such as a picture taken from your camera, or a scanned item does not contain data about the text within the item. Image-based media is merely an array of pixels. The rendering instructions for image-based media consists of pixel color and placing each pixel where it belongs in the image. This type of media is where the magic behind OCR comes into play. OCR is programmatically designed to “guess” the text within the image.

Modern OCR

OCR technology has been around since the early 90s, and therefore modern OCR is very effective in determining image text nowadays. Most products have >98% accuracy in determining text on a standard scanned document (300 dpi quality). Additionally, most OCR products support a variety of document types where the document text is encoded within the file.

How OCR is Utilized

Oftentimes, OCR is a hidden technology that lies within a product stack. End users are often unaware of the events that occur during the processing of a piece of media (referred to as a black box within the blow diagram).

The application flow works something like this—

  • Creating a search index for content repositories.
  • Assisting the blind and visually impaired.
  • Digitizing older texts (old books, newspapers, etc.) prior to the advent of digital media and writing.
  • Identifying text within media for metadata extraction (Ex. EZ Cloud and other Accounts Payable solutions).

Is your organization interested in automating Accounts Payable to optimize your existing process or capitalize on early pay discounts? EZ Cloud has a unique, patent-pending, and customizable extraction process, that both utilizes and enhances OCR to extract invoice metadata at a high degree of accuracy.

Contact EZ Cloud today for a demo that shows how our innovative process differs from the way organizations traditionally use OCR.