ABSTRACT
Detection of width changes in blood vessels of the retina may be indicative of eye or systemic disease. However, all fundus images, even those taken during a single sitting, are acquired at different perspective geometries, have different scales, and are difficult to manually compare side by side. This problem is increased for images acquired over time and is in no way lessened when different cameras are used as technology improves or patients change doctors. Indeed, even the method of image recording media has changed. In order to compare two images acquired at different times, automated techniques must be used to detect vessels, register images, and identify potential regions of blood vessel width change.
