Purchase this article with an account.
L. G. Epps, J. J. Hunter, C. J. Cookson, M. L. Kisilak, J. M. Bueno, M. C. W. Campbell; Quantifying the Quality of Confocal Scanning Laser Ophthalmoscopic Images of Blood Vessels. Invest. Ophthalmol. Vis. Sci. 2007;48(13):2755.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
High-resolution retinal imaging is vital to accurate disease detection and treatment monitoring. We used global image quality metrics and vessel contrast to quantify the effects on the quality of blood vessel images of controlled amounts of scattered light, used to model the potentially poorer quality associated with diseased and older eyes, and to optimize the confocal pinhole diameter across a broad age range.
A confocal scaning laser ophthalmoscope was modified to include a liquid crystal scatter generator in the scanning path, conjugated with the pupil of the eye. Video segments of the retina including blood vessels in the near periphery were recorded for differing degrees of scatter and no scatter in 12 healthy adults across a large range of ages (19-64 years). This was done for a set of confocal pinhole sizes. For each condition, frames were registered and averaged. For each averaged image, we calculated several overall metrics of image quality (signal to noise ratio, entropy and acutance) and the contrasts across one large and one small blood vessel. A repeated-measures ANCOVA (with age as the covariate) was performed for each image quality metric.
While entropy and acutance identified a significant (p<0.006) reduction in image quality with increasing amounts of induced light scatter, induced scatter had little effect on vessel contrast. All calculated contrasts were significantly (p<0.05) affected by the size of the confocal pinhole. Smaller pinholes generally resulted in higher contrast values, while the three overall metrics of image quality were comparably good with the three largest pinholes. Both contrast across the largest blood vessels and acutance were significantly affected by participant age, tending to decrease with increasing age. SNR did not reveal any significant effects.
Metrics of overall image quality are highly dependent on the amount of induced scatter while blood vessel contrast is more affected by the diameter of the confocal pinhole. A confocal pinhole corresponding to a 50µm projection onto the retina was found to be optimal for imaging blood vessels in people across a large range of ages and with induced scatter.
This PDF is available to Subscribers Only