Effect of Image Quality, Color, and Format on the Measurement of Retinal Vascular Fractal Dimension

Department of Ophthalmology, Centre for Vision Research, Westmead Millennium Institute, University of Sydney, Sydney, Australia
Investigative ophthalmology & visual science (Impact Factor: 3.4). 11/2010; 51(11):5525-9. DOI: 10.1167/iovs.09-4129
Source: PubMed


Fractal dimension of retinal vasculature is a global summary measure of retinal vascular network pattern and geometry. This study was conducted to examine the effect of variations in image color, brightness, focus, contrast, and format on the measurement of retinal vascular fractal dimension.
A set of 30 retinal images from the Blue Mountains Eye Study was used for a series of experiments by varying brightness, focus (blur), contrast, and color (color versus monochrome). The original and the modified images were graded for fractal dimension (D(f)) using dedicated retinal imaging software (IRIS-Fractal). A further set of 20 grayscale images was used to compare image format (.jpg versus .tif) with regard to the resultant D(f) and processing time.
The mean D(f) of original images in this sample was 1.454. Compared with the original set of images, variations in brightness, focus, contrast, and color affected the measurements to a small to moderate degree (Pearson correlation coefficient, r, ranged from 0.47 to 0.97). Very dark or blurry images resulted in a substantially lower estimate of D(f). Monochrome images were also consistently associated with lower D(f) compared with that obtained from color images. Using .jpg or .tif image formats did not affect the measurement or the time needed to process and measure D(f).
Variations in image brightness, focus, and contrast can significantly affect the measurement of retinal vascular fractals. Standardization of image parameters and consistent use of either monochrome or color images would reduce measurement noise and enhance the comparability of the results.

Download full-text


Available from: Alan Wainwright, Feb 24, 2015
  • [Show abstract] [Hide abstract]
    ABSTRACT: The present study examined the effects of blood pressure on a spectrum of quantitative and qualitative retinal microvascular signs. Retinal photographs from the Singapore Malay Eye Study, a population-based cross-sectional study of 3280 (78.7% response) persons aged 40-80 years, were analyzed. Quantitative changes in the retinal vasculature (branching angle, vascular tortuosity, fractal dimension, and vascular caliber) were measured using a semi-automated computer-based program. Qualitative signs, including focal arteriolar narrowing (FAN), arteriovenous nicking (AVN), opacification of the arteriolar wall (OAW), and retinopathy (e.g., microaneurysms, retinal hemorrhages), were assessed from photographs by trained technicians. After excluding persons with diabetes and ungradable photographs, 1913 persons provided data for this analysis. In multivariable linear regression models controlling for age, sex, BMI, use of antihypertensive medication, and other factors, retinal arteriolar branching asymmetry ratio, arteriolar tortuosity, venular tortuosity, fractal dimension, arteriolar caliber, venular caliber, FAN, AVN, and retinopathy were independently associated with mean arterial blood pressure. In contrast, arteriolar/venular branching angle, venular branching asymmetry ratio and OAW were not related to blood pressure. Retinal arteriolar caliber (sβ = -0.277) and FAN (sβ = 0.170) had the strongest associations with mean arterial blood pressure, and higher blood pressure levels were associated with increasing number of both quantitative and qualitative retinal vascular signs (P trend <0.001). Elevated blood pressure is associated with a spectrum of quantitative and qualitative retinal vascular signs, with the number of signs increasing with higher blood pressure levels.
    No preview · Article · Jul 2011 · Journal of Hypertension
  • [Show abstract] [Hide abstract]
    ABSTRACT: The main objective of this study is to determine a global index of the geometric complexity of the human retinal vascular network using the fractal geometry. Material and Methods: The fractal analysis of digital retinal images was performed with the Image J fractal analysis software and the fractal dimensions were calculated using the standard box-counting method. Results: The mean fractal dimension of normal retinal vascular network was 1.6147 ± 0.0151 in segmented variant and 1.5554 ± 0.0239 in skeletonised variant. Also, the mean lacunarity parameter was 0.5210 ± 0.0343 in segmented version and 0.2916 ± 0.0218 in skeletonised version.
    No preview · Article · Jan 2012 · Human and Veterinary Medicine
  • [Show abstract] [Hide abstract]
    ABSTRACT: Previous studies have reported high right-left eye correlation in retinal vessel caliber. We test the hypothesis that right-left correlation in retinal vessel caliber would be reduced in anisometropic compared with emmetropic children. Retinal arteriolar and venular calibers were measured in 12-year-old children. Three groups were selected: group 1, both eyes emmetropic (n = 214); group 2, right-left spherical equivalent refraction (SER) difference ≥1.00 but <2.00 diopter (D) (n = 35); and group 3, right-left SER difference ≥2.00 D (n = 32). Pearson's correlations between the two eyes were compared between group 1 and group 2 or 3. Associations between right-left difference in refractive error and right-left difference in caliber measurements were assessed using linear regression models. Right-left correlation in group 1 was 0.57 for central retinal arteriolar equivalent (CRAE) and 0.70 for central retinal venular equivalent (CRVE) compared with 0.60 and 0.82 for CRAE and CRVE, respectively, in group 2 (P = 0.42 and P = 0.08), and 0.36 and 0.52, respectively, in group 3 (P = 0.08 and P = 0.07, referenced to group 1). Each 1.00-D increase in right-left SER difference was associated with a 0.74-μm increase in mean CRAE difference (P = 0.02) and a 1.23-μm increase in mean CRVE difference between the two eyes (P = 0.002). Each 0.1-mm increase in right-left difference in axial length was associated with a 0.21-μm increase in the mean difference in CRAE (P = 0.01) and a 0.42-μm increase in the mean difference in CRVE (P < 0.0001) between the two eyes. Refractive error ≥2.00 D may contribute to variation in measurements of retinal vessel caliber.
    No preview · Article · Jul 2012 · Investigative ophthalmology & visual science
Show more