Segmentation of infected tissues in MRI based on VQ-BF filtering

Author

Graña, Manuel ; Echave, Imanol ; Ruiz-Cabello, Jesus ; Cortijo, Manuel

Author_Institution

EHU, Univ. Politecnica de Valencia, Spain

Volume

2

fYear

2002

fDate

26-30 Aug. 2002

Firstpage

1540

Abstract

We propose and test an hybrid procedure for the segmentation of infected tissues in MRI. The unsupervised step is the VQ-BF filtering of the MR volumes. It provides a robust preprocessing of the images, identifying basic texture elements regardless of image formation artifacts and geometric deformations. The supervised phase consist on the direct application of the empirical a posteriori distribution of the infection voxels on the VQ-BF images.

Keywords

Bayes methods; biological tissues; biomedical MRI; filtering theory; image coding; image segmentation; image texture; medical image processing; vector quantisation; MR volumes; MRI; VQ-BF filtering; a posteriori distribution; geometric deformations; hybrid procedure; image formation artifacts; image preprocessing; image texture elements; infected tissues segmentation; infection voxels; supervised phase; vector quantization-Bayesian filter; Animal structures; Bayesian methods; Filtering; Filters; Image coding; Image processing; Image restoration; Magnetic resonance imaging; Pixel; Vector quantization;

fLanguage

English

Publisher

ieee

Conference_Titel

Signal Processing, 2002 6th International Conference on

Print_ISBN

0-7803-7488-6

Type

conf

DOI

10.1109/ICOSP.2002.1180089

Filename

1180089