DISCOVERY

FLEX-Miner™

Search and Critical Word Spotting In Handwritten Documents

 

Just as handwriting examination is an art form dependent on the human eye, so is the ability to search images defined—and limited—by human effort. Visual searches are dependent on “metadata,” the descriptive text a person keys in and attaches to an imaged document, photo or video clip. Machine searches of imaged documents are also limited by OCR technology and until now, searching for foreign language words in handwritten content was next to impossible.

What if there was a way around all of that?
Gannon Technologies Group’s Graph-based Recognition™ technology structures physical data and its relational information in a very compact and highly computable structure. The technology employs principles of Graph Theory to produce a data structure in which individual pieces of data are organized within the framework of a mathematical graph. Graph-based Recognition™ technology shares common roots with traditional Optical Character Recognition (OCR), but takes it light years beyond traditional data capture. Graph-based Recognition™ technology retains much more of the information extracted from the original image, including characters, as well as data related to figure topology and geometry.

How Does it Work?
Graph-based Recognition™ technology detects the “signature” of words contained in images of documents. It begins with images of documents. The images are broken down into parts representing words, characters, and other items of interest. Each of these parts is converted into a graph containing a wealth of physical measurements and mathematical descriptors, rather like the architectural blueprint for a building. All of the measurements become the basis for decoding the graph into its actual identity as a character, a group of characters, or a word.

At the core of Graph-based Recognition™ is GTG’s Lossless Indexing, a technology that creates rich indexes for difficult to interpret content — handwritten texts, images, or cursive scripts. The better an index represents the content’s deeply embedded information and facets; the better a system can retrieve information, extract entities, and link events.

 

Topology & Features

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

FLEX-Miner™ is a trademark of Gannon Technologies Group