Detecting Blood Vessels in Eye image Using Fuzzy rubles

In this study, fuzzy neighbors rules are exploited to detect blood vessels in eye image. For this purpose a three level algorithm is utilized which includes preprocess, fuzzy system and post-process.A preprocess is needed to be performed in order to eliminate optical disk segment and FOV before applying fuzzy algorithm. Afterwards, vessels are detected using a fuzzy algorithm and presented rules for fuzzy set.Finally, applying complementary morphology operation improves segmentation. Images used in this article are chosen from eye image bank of DRIVE database including 10 retinal images.Obtained results demonstrate that proposed algorithm has suitable performance. The average of accuracy parameter of 10 images is derived 93%.

Keywords:

blood vessel , fuzzy logic , eye image , edge detection