Publication detail

Retinal Nerve Fiber Layer Analysis via Markov Random Fields Texture Modelling

ODSTRČILÍK, J. KOLÁŘ, R. HARABIŠ, V. GAZÁREK, J. JAN, J.

Original Title

Retinal Nerve Fiber Layer Analysis via Markov Random Fields Texture Modelling

Czech Title

Retinal Nerve Fiber Layer Analysis via Markov Random Fields Texture Modelling

English Title

Retinal Nerve Fiber Layer Analysis via Markov Random Fields Texture Modelling

Type

conference paper

Language

en

Original Abstract

The texture analysis of the retinal nerve fiber layer (RNFL) in colour fundus images is a promising tool for early glau-coma diagnosis. This paper describes model-based method for detection of changes in the RNFL. The method utilizes Gaussian Markov random fields (GMRF) and the least-square error (LSE) estimate for the local RNFL texture modelling. The model parameters are used as a texture fea-tures and non-linear classifier based on the Bayesian rule is used for classification of healthy and glaucomatous RNFL tissue. The proposed features are tested in the sense of clas-sification errors and also they are applied for segmentation of RNFL defects in high-resolution colour fundus-camera images. The results are also compared with the Optical Co-herence Tomography images regarded as a gold standard for our application due to the possibility of RNFL thickness measurement.

Czech abstract

The texture analysis of the retinal nerve fiber layer (RNFL) in colour fundus images is a promising tool for early glau-coma diagnosis. This paper describes model-based method for detection of changes in the RNFL. The method utilizes Gaussian Markov random fields (GMRF) and the least-square error (LSE) estimate for the local RNFL texture modelling. The model parameters are used as a texture fea-tures and non-linear classifier based on the Bayesian rule is used for classification of healthy and glaucomatous RNFL tissue. The proposed features are tested in the sense of clas-sification errors and also they are applied for segmentation of RNFL defects in high-resolution colour fundus-camera images. The results are also compared with the Optical Co-herence Tomography images regarded as a gold standard for our application due to the possibility of RNFL thickness measurement.

English abstract

The texture analysis of the retinal nerve fiber layer (RNFL) in colour fundus images is a promising tool for early glau-coma diagnosis. This paper describes model-based method for detection of changes in the RNFL. The method utilizes Gaussian Markov random fields (GMRF) and the least-square error (LSE) estimate for the local RNFL texture modelling. The model parameters are used as a texture fea-tures and non-linear classifier based on the Bayesian rule is used for classification of healthy and glaucomatous RNFL tissue. The proposed features are tested in the sense of clas-sification errors and also they are applied for segmentation of RNFL defects in high-resolution colour fundus-camera images. The results are also compared with the Optical Co-herence Tomography images regarded as a gold standard for our application due to the possibility of RNFL thickness measurement.

Keywords

glaucoma, retinal nerve fiber layer, retinal vessels segmentation, texture analysis, pattern recognition

RIV year

2010

Released

24.08.2010

Publisher

EURASIP

Pages from

1650

Pages to

1654

Pages count

4

BibTex


@inproceedings{BUT29969,
  author="Jan {Odstrčilík} and Radim {Kolář} and Vratislav {Harabiš} and Jiří {Gazárek} and Jiří {Jan}",
  title="Retinal Nerve Fiber Layer Analysis via Markov Random Fields Texture Modelling",
  annote="The texture analysis of the retinal nerve fiber layer (RNFL) in colour fundus images is a promising tool for early glau-coma diagnosis. This paper describes model-based method for detection of changes in the RNFL. The method utilizes Gaussian Markov random fields (GMRF) and the least-square error (LSE) estimate for the local RNFL texture modelling. The model parameters are used as a texture fea-tures and non-linear classifier based on the Bayesian rule is used for classification of healthy and glaucomatous RNFL tissue. The proposed features are tested in the sense of clas-sification errors and also they are applied for segmentation of RNFL defects in high-resolution colour fundus-camera images. The results are also compared with the Optical Co-herence Tomography images regarded as a gold standard for our application due to the possibility of RNFL thickness measurement.",
  address="EURASIP",
  booktitle="18th European Signal Processing Conference (EUSIPCO-2010)",
  chapter="29969",
  edition="EURASIP",
  howpublished="online",
  institution="EURASIP",
  year="2010",
  month="august",
  pages="1650--1654",
  publisher="EURASIP",
  type="conference paper"
}