Small retinal vessels extraction towards proliferative diabetic retinopathy screening

Authors:
Girish Singh Ramlugun;Vivek Krishna Nagarajan;Chandan Chakraborty
Affiliations:
Department of Biomedical Engineering, Manipal Institute of Technology, Manipal, Karnataka, India;Department of Biomedical Engineering, Manipal Institute of Technology, Manipal, Karnataka, India;School of Medical Science & Technology, IIT Kharagpur, India
Venue:
Expert Systems with Applications: An International Journal
Year:
2012

Citing 5
Cited 1

A Computational Approach to Edge Detection

IEEE Transactions on Pattern Analysis and Machine Intelligence
Realization of the Contrast Limited Adaptive Histogram Equalization (CLAHE) for Real-Time Image Enhancement

Journal of VLSI Signal Processing Systems
An improved matched filter for blood vessel detection of digital retinal images

Computers in Biology and Medicine
A modified matched filter with double-sided thresholding for screening proliferative diabetic retinopathy

IEEE Transactions on Information Technology in Biomedicine - Special section on biomedical informatics
Adaptive image contrast enhancement using generalizations of histogram equalization

IEEE Transactions on Image Processing

Evolutionary algorithm based classifier parameter tuning for automatic diabetic retinopathy grading: A hybrid feature extraction approach

Knowledge-Based Systems

Quantified Score

Hi-index	12.05

Visualization

Abstract

This paper introduces a systematic approach for segmentation of retinal blood vessels towards the diagnosis of proliferative diabetic retinopathy (PDR) from digital fundus images. Here, 2D-Match (Gabor) filters are used on digital fundus images in which the vessels are enhanced by contrast-limited adaptive histogram equalization (CLAHE). A double sided thresholding scheme is then used to segment the vessels. Hysteresis thresholding is performed with the large and small vessels clipped at different intensity levels in order to reconstruct the vessels in the image. This vessel extraction technique leads to high accuracy (93.1%) in comparison with the ground truth images provided in the DRIVE database.