Geometry-Driven-Diffusion Filtering of MR Brain Images using Dissimilarities and Optimal Relaxation Parameter

  • Ivan Bajla (Autor)
  • Igor Holländer (Autor)

Aktivität: Vortrag ohne Tagungsband / VorlesungPräsentation auf einer wissenschaftlichen Konferenz / Workshop


A novel method of local adapting of the conductance using a pixel dissimilarity measure is developed. An alternative processing methodology is proposed, which is based on intensity gradient histogram calculated for region interiors and boundaries of a phantom which models real MR brain scans. It involves a specific sot function suitable for the calculation of the optimum relaxation parameter Kopt and for the selection of the optimal exponential conductance. Computer experiments for locally adaptive geometry-driven-diffusion filtering of an MR brain phantom have benn performed and evaluated.
Zeitraum14 Juni 199817 Juni 1998
EreignistitelVIII Mediterranean Conference on Medical and Biological Engineering and Computing - Medicon '98

Research Field

  • Nicht definiert