Detection of Blood Vessels in Color Fundus Images using a Local Radon Transform

Document Type : Original Paper

Authors

1 PhD Student, Computer Engineering Dept., Ferdowsi University of Mashhad, Mashhad, Iran.

2 Associate Professor, Computer Engineering Dept., Ferdowsi University of Mashhad, Mashhad, Iran.

3 Assistant Professor, Ophthalmic Research Center, Khatam-Al-Anbia Hospital, Mashhad University of Medical Sciences, Mashhad, Iran.

Abstract

Introduction: This paper addresses a method for automatic detection of blood vessels in color fundus images which utilizes two main tools: image partitioning and local Radon transform.
Material and Methods: The input images are firstly divided into overlapping windows and then the Radon transform is applied to each. The maximum of the Radon transform in each window corresponds to the probable available sub-vessel. To verify the detected sub-vessel, the maximum is compared with a predefined threshold. The verified sub-vessels are reconstructed using the Radon transform information. All detected and reconstructed sub-vessels are finally combined to make the final vessel tree.
Results: The algorithm’s performance was evaluated numerically by applying it to 40 images of DRIVE database, a standard retinal image database. The vessels were extracted manually by two physicians. This database was used to test and compare the available and proposed algorithms for vessel detection in color fundus images. By comparing the output of the algorithm with the manual results, the two parameters TPR and FPR were calculated for each image and the average of TPRs and FPRs were used to plot the ROC curve.
Discussion and Conclusion: Comparison of the ROC curve of this algorithm with other algorithms demonstrated the high achieved accuracy. Beside the high accuracy, the Radon transform which is integral-based makes the algorithm robust against noise.

Keywords

Main Subjects

Iranian Journal of Medical Physics
Volume 7, Issue 3 - Serial Number 3
September and October 2010
Pages 1-8

Files

History

  • Receive Date: 28 February 2010
  • Revise Date: 18 July 2010
  • Accept Date: 27 June 2010

Share

How to cite

Statistics