Efficient and robust optic disc detection with vessel characteristics

A novel efficient and robust optic disc (OD) center detection method is proposed by using vessel distribution and directional characteristics. A feature describing local vessel density, compactness and uniformity of vessel distribution characteristics is presented to locate column ordinate of OD center. Then according to the global vessel direction characteristic, a General Hough Transformation (GHT) technology is introduced to identify the row ordinate of OD center. Four public datasets have been used to evaluate the proposed method. The OD was successfully located in 339 images out of 340 images (99.7%). And average computation times for STARE images are about 2.1-10.3s, which relate to image size. The proposed method shows excellent robustness on both normal and diseased images. It is better than many previous methods in the aspect of accuracy and efficiency.