Home > Archive > 2012 > Volume 2 Number 5 (Oct. 2012) >
IJMLC 2012 Vol.2(5): 593-597 ISSN: 2010-3700
DOI: 10.7763/IJMLC.2012.V2.196

Retinal Blood Vessel Segmentation Using Gabor Wavelet and Line Operator

Reza Kharghanian and Alireza Ahmadyfard

Abstract—In this paper, we propose a method for segmenting blood vessels from retinal images. We extract two sets of features for image classification: features based on Gabor wavelet and line operator. At each pixel of retinal image we construct a feature vector consisting of the pixel intensity, four features from Gabor wavelet transform in different scales and two features from orthogonal line operators. We compare the result of classification using two classifiers: Bayesian and SVM. First we estimate class-conditional probability density functions for vessel and non-vessel using Gaussian mixture model. Then using a Bayesian classifier we implement a fast classification. The result of experiments show the combination of Gabor features and line features provides a good performance for vessel segmentation. We tested the proposed algorithm on DRIVE database which is publicly available. As the second classifier we employ Support Vector Machine. The results shows SVM classifier in some cases performs better than Bayesian classifier.

Index Terms—Retinal image, vessel segmentation, Gabor wavelet, line detector, supervised classification.

The authors are with the Electrical and robotic engineering, Shahrood university of technology Shahrood, Iran (e-mail: r_kharghanian@yahoo.com; ahmadyfard@shahroodut.ac.ir).


Cite:Reza Kharghanian and Alireza Ahmadyfard, "Retinal Blood Vessel Segmentation Using Gabor Wavelet and Line Operator," International Journal of Machine Learning and Computing vol.2, no. 5, pp. 593-597, 2012.

General Information

  • ISSN: 2010-3700 (Online)
  • Abbreviated Title: Int. J. Mach. Learn. Comput.
  • Frequency: Bimonthly
  • DOI: 10.18178/IJMLC
  • Editor-in-Chief: Dr. Lin Huang
  • Executive Editor:  Ms. Cherry L. Chen
  • Abstracing/Indexing: Inspec (IET), Google Scholar, Crossref, ProQuest, Electronic Journals Library.
  • E-mail: ijmlc@ejournal.net

Article Metrics