Area Analysis of Foveal Avascular Zone in Diabetic Retinopathy Color Fundus Images
ABSTRACT Foveal avascular zone (FAZ) is the fovea devoid of capillaries in the macular region. An enlargement of FAZ is usually found in eyes with diabetic retinopathy (DR) resulting from a loss of capillaries in the perifoveal capillary network. Monitoring FAZ area enlargement at non-proliferative DR (NPDR) stage enables physicians to take preventive measures to curb progression of the disease to proliferative DR (PDR) stage and visual loss. Currently it is difficult to discern the FAZ area and to measure its enlargement in an objective manner using on color fundus images instead we require performing fundus fluorescein angiogram.
Fundus image analysis of FAZ presents several challenges such as high image, improper illumination, glare, fadeout, loss of focus and artifacts arising from reflection, refraction and dispersion. A new automated approach for area analysis of FAZ using color fundus images has been developed. Here, a binary map of retinal vasculature is computer generated from the color fundus image by extraction and reconstruction of retinal vasculature. The binary image is then used to determine vessel ends and pathologies surrounding FAZ for area analysis.
The performance of the FAZ area analysis method is compared to a manually-segmented FAZ from normal color fundus image. Digitally enhanced (CLAHE) red-free fundus images are used to establish segmented FAZ as a reference. The proposed method is found to achieve accuracies from 69.1% to 99.8% compared to accuracies of 16.6% to 98.9% for normal color fundus images. The proposed method is more consistent with mean accuracy improvement of 16.1%. The standard deviation achieved by the proposed method is also smaller by a factor of 2.8. The developed algorithm is thus more reliable than using color fundus image in the area analysis of FAZ.