Drusen detection from colored fundus images for diagnosis of age related Macular degeneration

Age related Macular degeneration (AMD) is an ocular disease caused by the extracellular material deposit at retinal pigment level known as drusen. AMD is the major cause of gradual vision loss. Early detection of the disease can be helpful to prevent blindness. Automatic fundus image analysis plays an important role for early diagnose of AMD. In this paper a novel method is presented for the detection of soft and hard drusen. After the preprocessing step the proposed method computes color and Gabor filter based features for each pixel to classify and extract all possible drusen pixels. Connected component labeling removes the suspicious pixels from drusen region and at the end optic disk is removed to avoid false positive pixels. The performance of proposed system was evaluated on publically available database STARE and it achieves 96% accuracy for drusen detection over STARE. The proposed algorithm shows significant improvement in detecting drusen for diagnosing AMD at earlier stage.