Fundus foveal localization based on vessel model.

The problem of automatically locating the fovea, a spot located in the center of the macula and responsible for sharp central vision, on the retinal surface image is considered. The uniqueness of the strategy presented here is the relatively small amount of computational expense of a quick and robust method conducive to real-time applications. The algorithm first identifies the main blood vessels using the modified active shape model (ASM) and then expeditiously attempts to fit the parabolic shape to the overall result to locate the fovea. A reliable tracking strategy is also proposed here to estimate the required vertex when the prior information of the optic disk is not available. When a patient's retinal vessel data have been modeled a priori and used as template, the technique brought up here should find its advantage in real-time image-guided applications like a computer-assisted photocoagulation treatment.