Publication detail

Computer support for early glaucoma diagnosis based on the fused retinal images

KOLÁŘ R., JAN J., KUBEČKA L.

Original Title

Computer support for early glaucoma diagnosis based on the fused retinal images

Czech Title

Počítačová podpora pro včasnou diagnostiku glaukomu založená na fůzovaných obrazech sítnice

English Title

Computer support for early glaucoma diagnosis based on the fused retinal images

Type

journal article

Language

en

Original Abstract

The autofluorescence of the retinal tissue has a potential to be used as a feature for the early glaucoma diagnosis [3]. In this paper, a new method for detection and segmentation of the zones with a higher level of autofluorescence around the optical disc in autofluorescent images is proposed. Areas of these small isles and their distances from the optical disc (OD) can be used as diagnostic parameters. The border of the OD is determined with the help of auxiliary infrared images, because of higher contrast between OD and surrounding tissue. Consequently, the multimodal registration method of both images to be fused must be applied, relying on our previous results [14]. The region growing method with manually determined seed points is used for segmentation. We show that the semi-automatic segmentation approach offers a reasonable estimation in comparison with manually segmented regions. The estimate is well defined thus removing the undesirable inter-expert variance.

Czech abstract

Článek se zabývá využitím autofluorescenčních (AF) obrazů sítnice, jako zdroje přídavných informací, pro včasnou diagnostiku glaukomového onemocnění. Je navržen postup pro detekci zón se zvýšenou autofluorescencí v okolí optického disku (OD).

English abstract

The autofluorescence of the retinal tissue has a potential to be used as a feature for the early glaucoma diagnosis [3]. In this paper, a new method for detection and segmentation of the zones with a higher level of autofluorescence around the optical disc in autofluorescent images is proposed. Areas of these small isles and their distances from the optical disc (OD) can be used as diagnostic parameters. The border of the OD is determined with the help of auxiliary infrared images, because of higher contrast between OD and surrounding tissue. Consequently, the multimodal registration method of both images to be fused must be applied, relying on our previous results [14]. The region growing method with manually determined seed points is used for segmentation. We show that the semi-automatic segmentation approach offers a reasonable estimation in comparison with manually segmented regions. The estimate is well defined thus removing the undesirable inter-expert variance.

Keywords

glaucoma, image segmentation, retinal images

RIV year

2006

Released

31.12.2006

Pages from

249

Pages to

259

Pages count

10

BibTex


@article{BUT44023,
  author="Radim {Kolář} and Jiří {Jan} and Libor {Kubečka}",
  title="Computer support for early glaucoma diagnosis based on the fused retinal images",
  annote="The autofluorescence of the retinal tissue has a potential to be used as a feature for the early glaucoma diagnosis [3]. In this paper, a new method for detection and segmentation of the zones with a higher level of autofluorescence around the optical disc in autofluorescent images is proposed. Areas of these small isles and their distances from the optical disc (OD) can be used as diagnostic parameters. The border of the OD is determined with the help of auxiliary infrared images, because of higher contrast between OD and surrounding tissue. Consequently, the multimodal registration method of both images to be fused must be applied, relying on our previous results [14]. The region growing method with manually determined seed points is used for segmentation. We show that the semi-automatic segmentation approach offers a reasonable estimation in comparison with manually segmented regions. The estimate is well defined thus removing the undesirable inter-expert variance.",
  chapter="44023",
  journal="Scripta medica",
  number="79",
  volume="2006",
  year="2006",
  month="december",
  pages="249--259",
  type="journal article"
}