A fully automated digital image processing system, which provides an o
bjective and repeatable way to quantify microaneurysms in digitised fl
uorescein angiograms, has been developed, The automated computer proce
ssing includes registration of same-eye retinal images for serial stud
ies, cutting of regions-of-interest centred on the fovea, the detectio
n of microaneurysms and the comparison of serial images for microaneur
ysm turnover, The microaneurysm detector was trained against a databas
e of 68 images of patients with diabetes containing 394 true microaneu
rysms, as identified by an ophthalmologist, The microaneurysm detector
achieved 82% sensitivity with 2.0 false-positives per image. An indep
endent test set, comprising 20 images containing 297 true microaneurys
ms, was used to compare the microaneurysm detector with clinicians, Th
e microaneurysm detector achieved a sensitivity of 82% for 5.7 false-p
ositives per image, whereas the clinician receiver-operator-characteri
stic (ROC) curve gives 3.2 false-positives per image at a sensitivity
of 82%, It is concluded that the computer system can reliably detect m
icroaneurysms, The advantages of the computer system include objectivi
ty, repeatability, speed and full automation.