Detail publikace

Segmentation methods for MRI processing

MIKULKA, J. GESCHEIDTOVÁ, E. BARTUŠEK, K.

Originální název

Segmentation methods for MRI processing

Typ

článek ve sborníku ve WoS nebo Scopus

Jazyk

angličtina

Originální abstrakt

The paper describes the segmentation of NMR image of the human head in the region of temporomandibular joint. Images obtained by the tomograph used are of very low resolution and contrast, and processing them can prove to be difficult. There were found many methods for segmentation of small and noised parts of NMR images such as active contours, region-based or edge-based level set segmentation methods and very known the region growing method. An appropriate algorithm has been found, which consists of pre-processing the image by a smoothing filter, sharpening, and four-phase level set segmentation. This method segments the image on the basis of the intensity of regions and is thus suitable to be applied to the above NMR images, in which no sharp edges occur. The method also has some filtering capability. The paper describes the most popular segmentation methods in MRI processing and it is shown the result of the mentioned four-phase level set segmentation method. The future work will be aimed at testing and comparing the results of each described methods and with manual or interactive segmentation in NMR images in the mentioned region of temporomandibular joint. The segmented slices will be used for 3D modeling of tissues.

Klíčová slova

level set segmentation, mandibular joint

Autoři

MIKULKA, J.; GESCHEIDTOVÁ, E.; BARTUŠEK, K.

Rok RIV

2009

Vydáno

23. 11. 2009

Místo

Polsko

ISBN

978-83-922095-8-4

Kniha

Recent Advances in Numerical Modelling

Strany od

147

Strany do

150

Strany počet

4

BibTex

@inproceedings{BUT31679,
  author="Jan {Mikulka} and Eva {Gescheidtová} and Karel {Bartušek}",
  title="Segmentation methods for MRI processing",
  booktitle="Recent Advances in Numerical Modelling",
  year="2009",
  pages="147--150",
  address="Polsko",
  isbn="978-83-922095-8-4"
}