A Radon Transform Based Approach for Extraction of Blood Vessels in Conjunctival Images

Authors:
Reza Pourreza;Touka Banaee;Hamidreza Pourreza;Ramin Daneshvar Kakhki
Affiliations:
Electrical Engineering Department, Ferdowsi University of Mashhad, Iran;Ophthalmic Research Center, Khatam-Al-Anbia Hospital, Medical Sciences University of Mashhad, Iran;Computer Engineering Department, Ferdowsi University of Mashhad, Iran;Ophthalmic Research Center, Khatam-Al-Anbia Hospital, Medical Sciences University of Mashhad, Iran
Venue:
MICAI '08 Proceedings of the 7th Mexican International Conference on Artificial Intelligence: Advances in Artificial Intelligence
Year:
2008

Citing 4
Cited 1

A review of vessel extraction techniques and algorithms

ACM Computing Surveys (CSUR)
Blood Vessel Detection via a Multi-window Parameter Transform

CBMS '06 Proceedings of the 19th IEEE Symposium on Computer-Based Medical Systems
Multiscale centerline extraction of angiogram vessels using gabor filters

CIS'04 Proceedings of the First international conference on Computational and Information Science
Accurate Centerline Detection and Line Width Estimation of Thick Lines Using the Radon Transform

IEEE Transactions on Image Processing

A complementary method for automated detection of microaneurysms in fluorescein angiography fundus images to assess diabetic retinopathy

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper proposes a local Radon transform-based algorithm for extraction of blood vessels in conjunctival images. This algorithm divides the image into overlapping windows and applies Radon transform to each window. Vessel direction in each window is found by detection of peak in Radon space. The proposed algorithm is capable of extracting blood vessels with a variety of widths. According to vessel width, extracted blood vessels are classified into some predefined classes and several statistics are computed for each class. Since the Radon transform is robust against noise, proposed algorithm is noise-independent and is more robust in comparison with other available algorithms.