Click here for Top of Page
Right Arrow: Next
Right Arrow: Previous

Global Pattern Recognition Series:

German Research Center for

Artificial Intelligence GmbH (DFKI)

Founded in 1988, DFKI today is the largest nonprofit contract research institute in the field of innovative software technology based on Artificial Intelligence (AI) methods. DFKI focuses on the complete cycle of innovation—from world-class basic research and technology development through leading-edge demonstrators and prototypes to product functions and commercialization. An important element of DFKI's mission is to move innovations as quickly as possible from the lab into the marketplace. Only by maintaining research projects at the forefront of science can DFKI have the strength to meet its technology transfer goals.

Currently, the DFKI GmbH employs 244 highly skilled people. More than 200 student research assistants support them on a part time basis. With an overall annual budget in 2006 of more than 21.0 million, the previous year’s record results of 18.1 million were surpassed and once again a positive annual net profit was reported. The circle of DFKI industrial shareholders includes DaimlerChrysler AG, Deutsche Telekom AG, SAP AG, IDS Scheer AG, Bertelsmann AG, and Microsoft Deutschland GmbH, and has recently been expanded by Deutsche Post World Net (DPWN) and BMW AG. DFKI is involved in "Igniting Ideas!", the high-tech strategy of the German Government. DFKI has created 1200 new Jobs in IT industry and 47 start-up and spin-off companies. The company is a hot spot for academic talent, 49 former DFKI employees are now professors in 11 different countries.

Pattern Recognition research and technology development is carried out at two of the DFKI research labs: Image Understanding and Pattern Recognition (IUPR) and Knowledge Management (KM).

The IUPR research lab at DFKI conducts basic and applied research in pattern recognition, machine learning, image understanding, and artificial intelligence, with practical applications to digital libraries, network security, bioinformatics, historical document analysis, and scientific data analysis.  Current significant projects include the IPeT project, which aims to create easy-to-use libraries and tools for integrating image analysis technologies into personal computing environments; as part of the IPeT project, we have developed image matching, OCR, and image retrieval libraries and servers, and demonstrated applications to image-based spam detection, analyzing and mediating accessibility, and automatic usability analysis.  The Netshield project is concerned with applying pattern recognition and machine learning to identifying and automatically defending against network attacks.  The OCRopus project aims to apply state-of-the-art techniques in machine learning and pattern recognition to optical character recognition for massive document conversion effort (e.g., Google Books).  Other application areas include automated detection of forged documents, historical document analysis, and camera-based document interaction.

Today, the KM research lab at DFKI conducts mainly applied research in text- and image-based document analysis, information retrieval and information extraction. The current project HyperPrinting aims at a new paper-based interaction paradigm, allowing the user to interact with the computer by means of annotations on the printed pages of a document. Therefore specific markers indicating well defined commands are overlaid on the print image. The generated markers are related to logical elements which in turn are determined by analysis specialists. These logical elements are e.g. title, author, abstract, and bibliography on a scientific publication but also more complex objects such as tables. The project strongly relies on technologies that were developed in more basic research related former projects of the KM department. One of these technologies is T-Recs++, a model free table recognition and structure analysis system which neither requires table delineation nor column spacing nor known column headers.

Feature Articles on uses of

Pattern Recognition (PR)



PR in Digital Libraries, Jul. ‘06

             [html]     [pdf]


PR at the US Postal Service:  A Decade of Achievement, Apr. ‘06

             [html]     [pdf]


PR in Traffic Engineering, Jul. ‘05

             [html]     [pdf]


PR in Astronomy and Photonics, Apr. ‘05

             [html]     [pdf]


PR in Origami, Jan. ‘05

             [html]     [pdf]


PR in Defense Applications, Jan. ‘04



PR in Maps, Sep. ‘03



PR in Security and Entertainment, Jun. ‘03



PR in Sports, Apr. ‘03


Global Pattern Recognition:


China’s National Laboratory of Pattern Recognition (NLPR)       

             [html]     [pdf]


PR in Two National Labs, Jan. ‘06

             [html]     [pdf]

By Andreas Dengel


Professor of Computer Science, University of Kaiserslautern
Scientific Director and Member of the Board, German Research Center for Artificial Intelligence (DFKI)