It is the object of a method and apparatus for retinal vessel analysis based on digital images to enhance the ability to discriminate between healthy vessel states and at-risk vessel states while reducing manual effort and saving time in order to allow individual vascular risk, particularly stroke risk, to be determined in a more reliable manner and with fewer subjective systematic and random errors. The vessel segment diameter, type of vessel and the image coordinates are determined for a series of adjoining vessel segments along vessel portions in a measurement zone surrounding the papilla and are stored by vessel segment with reference to the evaluated image, to a reference image recorded with a time offset, and to a displacement vector that is determined for the vessel segment between the reference image and an evaluated comparison image. Comparison measurements are carried out only on identical vessel segments already measured in the reference image. The correlation of vessel segments to vessel portions and to vessel type is adopted intact from the reference image. The stored data sets for the vessel segments of the reference image and comparison images provide a progression of coordinate-oriented vessel segment diameters for all measured vessel segments as basis for determining parameters and presenting them in a spatially resolved progression, e.g., in progress images.