Automated Optic Disc removal in fundus images using iterative heuristics and morphological operations

Abstract

Automatic Optic Disc (OD) removal is vital step prior to screening Diabetic Retinopathy (DR) cases. Hence in this paper, we proposed an automated technique for removal of OD position in digital retina fundus images based on mathematical morphological operations. Here, dilation and erosion to remove OD and enhance the appearance of binary images. The proposed technique was tested on 400 images of the publicly available MESSIDOR database. Experimental Results showed that OD can be effectively be removed from the binary images and the proposed method is robust too.