OCR

The ImageGear Recognition component provides access to document recognition technology.

This core software module provides recognition for almost any document, including those produced on typewriters, dot-matrix printers, ink-jet printers, laser printers, and phototypesetters, as well as photocopied and faxed versions of any document.

Using the ImageGear Recognition component, you can recognize a document image and export the text in one of the supported output formats. These formats include 8-bit ASII and 16-bit Unicode text, Microsoft Word, HTML, and others.

The ImageGear Recognition Component also allows a user to:

Manage delineated zones of a document page and then specify treatment for those zones. This includes the ability to correct the OCR engine's automatic segmentation between the segmentation phase and the recognition phase.
Process both text and graphics. The recognition software's ability to distinguish graphics from text can provide the basis of a compound document processing system.
Automatically detect fax, dot matrix, and other degraded documents and compensate accordingly.
Recognize Chinese Simplified, Chinese Traditional, Japanese, and Korean languages.

Portions of this document are excerpted from OmniPage Capture SDK 18.6 reference materials. Copyright © 2011 Nuance Communications, Inc. All Rights Reserved. Nuance and the Nuance logo are trademarks or registered trademarks of Nuance Communications, Inc. or its affiliates in the United States and/or other countries.

The figure below illustrates a simple application that takes a page image as input and produces recognized text in a variety of output formats.

ImageGear's Recognition component enables you to build OCR applications for the Windows development environment. An application using the Recognition API can:

Accept as input any image file format supported by ImageGear. This image data can be binary, gray, or color.
Recognize an image.
Save recognition data in either a simple text format, or one of document formats, such as MS Word or XML. The component supports the saving of multi-page documents.
Select the language or languages of documents to be recognized. A total of 123 languages, including the following:
- Latin alphabet languages, such as English, Eastern and Western European languages, and Baltic languages
- Cyrillic alphabet languages, such as Russian
- Asian languages, such as Chinese, Japanese, and Korean
- Turkish
- Greek
Documents with multiple languages can be recognized with accuracy because the API allows the application to specify the set of languages for recognition.

Enable end users to verify text during the recognition process.
Increase recognition accuracy with built-in and user-defined dictionaries.
Output confidence values for post-recognition processing.
Automatically segment the page to correctly recognize text on pages with complex or irregular layouts, including tables and graphics as well.
Allow the user to delineate zones of a document page and then specify treatment for those zones. This includes the ability to correct the OCR engine's automatic segmentation between the segmentation phase and the recognition phase.
Process both text and graphics. The recognition software's ability to distinguish graphics from text can provide the basis of a compound document processing system.
Automatically detect fax, dot matrix, and other degraded documents and compensate accordingly.
Use the scalable voting architecture that provides developers with two pre-made voting interfaces (PLUS2W, PLUS3W) and direct access to OCR engines.

This section provides information about the following: