Regensburg Medical Image Computing (ReMIC) Laboratory

Mission statement: "All primal, basic thought happens in pictures:
that is why the imagination is such a necessary tool, and why unimaginative brains will never achieve any greatness – except in mathematics"
(Arthur Schopenhauer, 1788-1860)
Purpose:

The Regensburg Medical Image Computing (ReMIC) laboratory is concerned with the computer-aided analysis and visualizing of medical image data. Our students learn the latest methods of image processing, starting with a fundamental knowledge of convolutions, smoothing and edge recognition, through Fourier transform and morphological image processing to complex algorithms for segmenting, registering and texture analysis.

They process data that are generally of varying origin (modality), size and resolution. A thermographic camera with built-in photo unit provides first-hand experience of images of different modality. Trends in medical imaging are moving towards 3D image data, whether native from computer tomography (CT), magnetic resonance tomography (MRT) and positron emission tomography (PET), or reconstructed from a stack of 2D images. To visualize this high-dimensional image data, the laboratory provides a VR wall with active stereo that enables the data to be viewed in genuine 3D. 3D render techniques from the lectures can be implemented and experienced first-hand here.

By engaging with topical aspects of current research and development projects the students develop a feeling for the needs and expectations of their medical partners and acquire a sense of the scope and limitations of automated image analysis.

In addition to teaching, the laboratory is designed to be a research facility where internal Bachelor and Master projects and – in future – joint doctoral dissertations can be worked on.[1]) Developing graphics algorithms for image processing and for the use of GPGPU is a particular area of interest, and the laboratory provides a GPU supercomputer for this purpose.
Learning objectives: 
  • Use of current methods of medical image processing
    • Convolution, Fourier transform, morphological image processing
    • Segmenting, registration, texture analysis
  • Foundation 3D visualizing
    • Ray tracing, ray casting
  • In-depth problem analysis and an imaginative use of image processing tools
  • Continuing development of known image processing algorithms
  • Recognising the opportunities and limitations of automated image analysis
 
Research objectives:

The overall aim of research done at the Regensburg Medical Image Computing (ReMIC) Laboratory is to create 3D models from image data of diverse origin, size and resolution.

Current research projects and corresponding publications will be found on the external homepage of   ReMIC, also current Invitations to render for Bachelor and Master dissertations [only available in German].

[1] German universities of applied sciences do not award doctoral degree themselves but can run doctoral programmes jointly with a partner university which awards the doctoral degree