Print Univis Search
FAU-Logo
Techn. Fakultät Willkommen am Institut für Informatik FAU-Logo
Sickel

Dipl.-Inf. Konrad Sickel

Researcher in the Medical Image Processing group at the Chair of Computer Science 5 (Pattern Recognition)
of the Friedrich-Alexander University Erlangen-Nuremberg

Research on a system which uses expert knowledge and improve itself through experience.

Projects

Project 01: Development of an expert system to support the editing of 3D surface models

Abstract:

During the production of medical prosthesis it is often necessary to customize the product for a patient. Today this is done by skilled operators using CAD software. The virtual models are produced by rapid prototyping techniques like 3D-Printing or stereolithography. The manual labor influence has several disadvantages. First of all the result differs from operator to operator and it is not reproducible.

To avoid these disadvantages an expert system shall be designed. First step is to acquire the expert knowledge of the operators and represent it. The knowledge can then be used by the inference machine of the EPS to make suggestion to the operators what to do next. The rules of the knowledge base are mainly based on patterns / feature points defined by written work instructions and interviews of experts.

Implementation of the expert system prototype will be done with the free expert system building tool CLIPS. To use CLIPS has several advantages. The tool is completely open source and can be used in every software system. It supports partitioning of the knowledge base in modules for easier maintenance and object orientated knowledge bases. The inference engine uses the Rete-Algorithm which is the standard for EPS shells at the moment.

 

Sheme of an expert system

Project 02: Development of an learning component for an expert system to automate the editing of 3D surface models

Abstract:

Goal of this project is to improve the developed expert system of project 01. The stored information of the interaction between user and EPS shall be used to modify the initial knowledge base. This modification shall be done by applying machine learning algorithms (CN2, FOIL, ...). Aim is to improve the quality of the output as well as create an interface to automatically learn new rules if new components are introduced.

 

Supported by:

Siemens AG in close cooperation with Siemens Corporate Research (Princeton, NJ)