Retinal blood vessels extraction from fundus images using an automated method

In the present-day, intuitive blood vessels finding is an indispensable task for identification of copious eye anomalies. So, this paper presents an instinctive and fast process for detection of blood vessels from fundus images. In this scheme the input image is primarily pre-processed by means of contrast limited adaptive histogram equalization (CLAHE) to enhance the blood vessels and then an optic disk removed image is obtained by subtracting the morphologically opened image and enhanced image. The blood vessels are then take out using ISODATA technique. To end a morphological cleaning action is applied to acquire the ultimate segmented image. The performance of the proposed method is assessed by means of three publicly offered DRIVE, STARE and CHASE_DB1 databases and attains average accuracies of 0.946, 0.949 and 0.948 on DRIVE, STARE and CHASE_DB1 databases respectively.