A new algorithm to detect the retinal structures

This paper deals with a new algorithm for automatic edge detection of retinal structures on angiographic frames. Usually, the determination of these structures and their accurate locations are based on a manual mapping wich is time consuming, tedious and not error-free. Computer image processing allows to automate the mapping and increases the reliability of extracted parameters. Therefore, we realize a process to automatically detect retinal structures. This process is based on the photometric and geometric properties of these structures. The results shown that the automatic mapping is more accurate and more reliable than manual mapping.