Chapter

Bildverarbeitung für die Medizin 2000

Part of the series Informatik aktuell pp 116-120

Elastic Distortion of Deformable Feature Maps for Fully-Automatic Segmentation of Multispectral MRI Data Sets of the Human Brain

  • Axel WismüllerAffiliated withInstitut für Radiologische Diagnostik, Ludwig-Maximilians-Universität München
  • , Frank VietzeAffiliated withInstitut für Radiologische Diagnostik, Ludwig-Maximilians-Universität München
  • , Dominik R. DerschAffiliated withCrux Cybernetics Corp
  • , Gerda LeinsingerAffiliated withInstitut für Radiologische Diagnostik, Ludwig-Maximilians-Universität München
  • , Johannes BehrendsAffiliated withInstitut für Radiologische Diagnostik, Ludwig-Maximilians-Universität München
  • , Helge RitterAffiliated withAG Neuroinformatik, Universität Bielefeld
  • , Klaus HahnAffiliated withInstitut für Radiologische Diagnostik, Ludwig-Maximilians-Universität München

* Final gross prices may vary according to local VAT.

Get Access

Abstract

In this paper, we present an algorithm that provides adaptive plasticity in function approximation problems: the deformable (feature) map (DM) algorithm. The DM approach reduces a class of similar function approximation problems to the explicit supervised one-shot training of a single data set. This is followed by a subsequent, appropriate similarity transformation which is based on a self-organized deformation of the underlying multidimensional probability distributions. After discussing the theory of the DM algorithm, we present results of its application to the real-world problem of fully automatic voxel-based multispectral image segmentation, employing magnetic resonance data sets of the human brain.

Keywords

Algorithmen Magnetresonanz Segmentierung Neuronale Netze Vektorquantisierung