An Approach to Context-driven Document Analysis and Understanding

Most of today‘s document analysis systems are employed for specialpurpose tasks and carry their knowledge directly within their algorithms. However, certain office environments demand for more flexible applications of document analysis and understanding (DAU) techniques having a broader application space and being subject to frequent changes while still requiring high precision. Our knowledge-centered approach to that represents not only comparatively static knowledge about document properties and analysis results within the same declarative formalism, but also includes the analysis task and the current system context within this formalism. This allows an easy definition of analysis tasks and an efficient and accurate analysis by using expectations about incoming documents as context information. Our approach has been implemented within the VOPR1 system which gains the required context from a commercial workflow management system (WfMS) by constant exchanges of expectations and analysis tasks.