Deutsch Intern
    Chair of Computer Science VI - Artificial Intelligence and Applied Computer Science

    Projects

    The Chair of Artificial Intelligence and Knowledge Systems is currently involved in a variety of projects. In the following, you will find information on current research projects:

    Image Processing:

    Segmentation of old prints and integration into OCR pipeline: Recognition of image and text regions in scans of old prints using different methods (including pixel classifier, contour-based approaches, baseline approaches), recognition and interpretation of tables in documents. Recognition of handwritten artifacts in old scans (project: internal project with ZPD, Würzburg; successor of joint project Kallimachos).

    Image analysis in medicine: Detection of polyps in real time in endoscopy videos (Project: IZKF with Prof. Meining), interpretation of radiological and endoscopic images, localization of an endoscope in the intestine.

    OMR (Optical Music Recognition): Transcription of song texts in old manuscripts: Segmentation of notes and text as well as their transcription and assignment (Project: Corpus Monodicum with Prof. Haug)

     

    Natural Language Processing:

    Legal judgment analysis to search for similar judgments to a legal case (Jana project with RenoStar GmbH)

    Medical information extraction from textual documents such as diagnostic reports, request documents (DZ-PTM project with University Hospital Würzburg and Prof. Hotho, Prof. Latoschik)

    Character Recognition in Novels (Succession of Kallimachos)

    Data extraction from tables in documents

    Chatbots for ethical and legal topics

     

    Knowledge processing and exploration:

    Simulation of heating systems for anomaly detection and optimization (KINERGY project of BMWi with ENER-IQ, SEnerCon, and Prof. Seipel)

    Development of a data warehouse with exploration of data and ad-hoc information extraction (focus: medical applications)

     

    Tutoring systems:

    E-assessment for semi-structured tasks, including tasks on programming (Python) and on web programming and data structures (databases, XML, class diagrams) Project  IT4all (Digital Campus Bavaria)

     

    Knowledge-based systems and logic programming:

    We are developing the declarative toolkit Declare for knowledge-based
    systems in SWI Prolog.

    Advanced applications require the management and extensive analysis
    of hybrid information (with incompleteness and uncertainty),
    e.g. diagnosis  and association rule mining in sensor data.
     
    Declare contains - amoung others - a deductive database system DDBase,
    a subsystem DisLog for reasoning in deductive databases (with
    disjunctivions), a language FNQuery for querying and transforming
    XML databases, and a graphical user interfaces in the object-oriented
    extension XPCE of SWI Prolog.