Abstract
The aim of the work was to investigate automatic methods of assessing the image quality of retinal images from premature infants taken with the RetCam 120 digital fundus camera. The quality of 73 images was assessed by measuring the asymmetry of histograms derived from edge maps. The edge maps were generated using a Gaussian matched filter, a method optimized for the detection of vessels on retinal images. The results show that measurement of the histogram asymmetry derived from edge maps is a promising method for the automated quantification of image quality of retinal images from infants.