PRESS RELEASES

Gannon Technologies Group files new patent application

Technology Breakthrough Enables Machine Understanding of Arabic Handwritten Documents

(January 2006)

 

Alexandria, VA – January 6, 2006 - Gannon Technologies Group, the leader in graph theory based pattern recognition, announces filing of a patent application for Pictographic Recognition Technology Applied to Distinctive Characteristics of Handwritten Arabic Text. 

GTG’s graph theory based – or “Pictographic” – pattern recognition technology converts handwritten documents to a rich representation of the observed graphical features. These features are used as the foundation for identifying written words as well as identifying the writers.  This patented technology is the basis for GTG’s search product – Flex-Miner- and handwriting biometric product – Flex-Tracker.

For Arabic handwriting, GTG’s FLEX-Miner technology employs the unique patent pending approach announced today.   The approach utilizes Parts of Arabic Words (PAWs) - individual characters, groups of characters or sometimes entire words. PAWs are connected cursive strings separated by white space. They are artifacts of the property of written Arabic that certain characters never connect to the following character. PAWs represent a breakthrough in Arabic handwriting recognition because they account for the high degree of “license” that individual writers take by omitting characters and diacritical markings as well as writing in a very stylized manner.

FLEX-Miner renders handwritten documents into indices of PAWs that can be used to search for specified key words. In addition to recognizing key words, the system can analyze the unique style of the writer.

 

Interested in learning more about GTG’s groundbreaking work?
Contact us at info@gannontech.