April 2010
Volume 51, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2010
Comparison of Retinal Nerve Fiber Layer Defect Evaluation Between RGB Signal Values of Color Fundus Photograph and Signal Value of Scanning Laser Ophthalmoscope
Author Affiliations & Notes
  • T. Sone
    Ophthalmology & Visual Science,
    Hiroshima University, Hiroshima, Japan
  • K. Kaneda
    Department of Information Engineering,
    Hiroshima University, Hiroshima, Japan
  • A. Hirota
    Hirota Eye Clinic, Yamaguchi, Japan
  • M. Takamatsu
    Ophthalmology & Visual Science,
    Hiroshima University, Hiroshima, Japan
  • H. Nakamoto
    Hirota Eye Clinic, Yamaguchi, Japan
  • Y. Kiuchi
    Ophthalmology & Visual Science,
    Hiroshima University, Hiroshima, Japan
  • Footnotes
    Commercial Relationships  T. Sone, None; K. Kaneda, None; A. Hirota, None; M. Takamatsu, None; H. Nakamoto, None; Y. Kiuchi, None.
  • Footnotes
    Support  None.
Investigative Ophthalmology & Visual Science April 2010, Vol.51, 2744. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      T. Sone, K. Kaneda, A. Hirota, M. Takamatsu, H. Nakamoto, Y. Kiuchi; Comparison of Retinal Nerve Fiber Layer Defect Evaluation Between RGB Signal Values of Color Fundus Photograph and Signal Value of Scanning Laser Ophthalmoscope. Invest. Ophthalmol. Vis. Sci. 2010;51(13):2744.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: : To examine the detection ability of retinal nerve fiber layer defect (NFLD) statistically from RGB signal value of the color fundus photography and signal value of SLO.

Methods: : Seventeen eyes (13 patients) that the retinal nerve fiber layer defect were visible on the stacked fundus images with a 532nm green source of scanning laser ophthalmoscope (SLO, F-10®, Nidek) were enrolled. Using the stacking method which processes the 12 - 15 fundus images taken at the same angle, F-10® can generate the monochrome fundus images with higher contrast than conventional SLO. Generally, each pixel on the digital photograph can have 256 levels of red (R), 256 levels of green (G), and 256 levels of blue (B). We made the original software which can calculate the R, G, B signal values of the points on the photograph, and compared the values at inside and outside of the NFLD on the fundus color photographs (2.0 mega pixels, JPEG format, Vx-10i®, Kowa) and SLO images statistically. The greater the signal value differentiation is, the higher the contrast is.

Results: : In all images, NFLD was seen more clearly on SLO images than on the color photographs visually. The mean signal values of the points at outside and inside of the NFLD are 155.6 ± 25.0 (outside) / 108.4 ± 23.2 (inside) on SLO, and R value = 184.4 ± 29.1 / 177.4 ± 32.3, G = 122.1 ± 24.2 / 113.2 ± 27.0 and B = 52.4 ± 22.4 / 42.8 ± 21.4 on the color fundus photographs respectively. The mean difference of absolute values are R = 7.0 ± 11.6, G = 8.9 ± 8.3 and B = 9.6 ± 6.1, and the average of relative values are 0.960 ± 0.06, 0.921 ± 0.074 and 0.775 ± 0.163 respectively. It can be considered that the highest contrast can be on the blue factor among R, G, and B on NFLD observation. The mean signal value on SLO is 47.2 ± 20.4, and it is significantly higher than R, G and B value (P<0.001). The average of relative values of SLO is 0.699 ± 0.117 and there is no significant difference with B value (P=0.12).

Conclusions: : Using R, G and B value, the findings on images can be analyzed by not only visual evaluation but also statistical calculation. SLO can visualize NFLD with higher contrast than color photograph statistically. On the color photograph, B signal is important to visualize NFLD.

Keywords: imaging/image analysis: clinical 
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×