Share Email Print

Proceedings Paper

Novel algorithm by low complexity filter on retinal vessel segmentation
Author(s): Samad Rostampour
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

This article shows a new method to detect blood vessels in the retina by digital images. Retinal vessel segmentation is important for detection of side effect of diabetic disease, because diabetes can form new capillaries which are very brittle. The research has been done in two phases: preprocessing and processing. Preprocessing phase consists to apply a new filter that produces a suitable output. It shows vessels in dark color on white background and make a good difference between vessels and background. The complexity is very low and extra images are eliminated. The second phase is processing and used the method is called Bayesian. It is a built-in in supervision classification method. This method uses of mean and variance of intensity of pixels for calculate of probability. Finally Pixels of image are divided into two classes: vessels and background. Used images are related to the DRIVE database. After performing this operation, the calculation gives 95 percent of efficiency average. The method also was performed from an external sample DRIVE database which has retinopathy, and perfect result was obtained

Paper Details

Date Published: 30 September 2011
PDF: 6 pages
Proc. SPIE 8285, International Conference on Graphic and Image Processing (ICGIP 2011), 82851C (30 September 2011); doi: 10.1117/12.913411
Show Author Affiliations
Samad Rostampour, Islamic Azad Univ. (Iran, Islamic Republic of)

Published in SPIE Proceedings Vol. 8285:
International Conference on Graphic and Image Processing (ICGIP 2011)
Yi Xie; Yanjun Zheng, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?