HOME ROSITEMAP

Case study: Intelligent Document Recognition

 

 Field of expertise: Image Processing


The client: Formic


Formic is one of the top 5 European information capture software vendors* and part of the Stockford Group of over 40 thriving companies in the US and UK, which is owned and backed by Sir Peter Michael CBE. Formic provides class-leading data capture software products that process millions of critical business forms daily, reducing the cost of delivering information to enterprise systems.

 

 

Business Overview


Our client’s software solution reads the information contained in documents whether on paper, the Internet, tablet PCs or PDAs then subjects each piece of data to a series of rigorous checks before transferring it to different enterprise applications. It also stores an image of the original document alongside the extracted data, for instant retrieval.

EvoSoftware has developed software solutions for Intelligent Document Recognition, designed to process scanned forms, identify various types of documents and extract the contained data.

 

Technical Overview


At first we have developed a pilot project enhancing the library functionality with features for line detection and line removal, edge detection as well as skew detection and automatic deskewing of black & white image documents. Some other functionality was in the template registration followed with features such as logo detection, template recognition and much more.

Currently we are addressing a text detection project having full functionality that includes paragraph and table structures detection.

 

Project challenge


Designing image processing algorithms that works at very high execution speeds yielding accurate results for a wide set of input documents.

 

Technology


C/C++ and ATL/COM on Microsoft OS platforms