Purpose:
To estimate RNLT(retinal nerve fiber layer thickness) from color fundus photograph
Methods:
After taking the fundus color photograph with a digital fundus camera offering 5.3 million pixels as the standard Windows® bitmap file format, we created the computer software for the image data processing. The software can calculate the R(Red), G(Green), B(Blue) and GB(Green-Blue; red free) color values on the peripapillary circle as CircleR, CircleG, CircleB and CircleGB and draw the graph of these values (Figure). To examine which values were correlated most with RNLT, the peripapillary RNLT was evaluated with RNLT (3.4) protocol on the Zeiss Stratus OCT system (Carl Zeiss Meditec, Inc., Dublin, CA). The CircleR, CircleG, CircleB, CircleGB were compared with the RNLT in each of 12 clock-hour sectors. 30 glaucoma suspect eyes of 20 patients were studied. The eyes with the peripapillary atrophy and large disc, and the defective photograph were excluded.
Results:
The 20.0% eyes of CircleR, 26.7% of CircleG, 46.7% of CircleGB and 93.3% of CircleB is correlated with RNLT. On individual case, the highest correlation with RNLT were CircleB(86.7% eyes) and CircleGB(13.3% eyes).
Conclusions:
The blue value of RGB may be the most important to estimate RNLT on the color fundus photograph.
Keywords: imaging/image analysis: clinical • nerve fiber layer • detection