A fast, efficient and automated method to extract vessels from fundus images

Abstract

We present a fast, efficient, and automatic method for extracting vessels from retinal images. The proposed method is based on the second local entropy and on the gray-level co-occurrence matrix (GLCM). The algorithm is designed to have flexibility in the definition of the blood vessel contours. Using information from the GLCM, a statistic feature is calculated to act as a threshold value. The performance of the proposed approach was evaluated in terms of its sensitivity, specificity, and accuracy. The results obtained for these metrics were 0.9648, 0.9480, and 0.9759, respectively. These results show the high performance and accuracy that the proposed method offers. Another aspect evaluated in this method is the elapsed time to carry out the segmentation. The average time required by the proposed method is 3 s for images of size 565 9 584 pixels. To assess the ability and speed of the proposed method, the experimental results are compared with those obtained using other existing methods.

Retinal Vessels Segmentation Techniques and Algorithms: A Survey

“…A notable vessel extraction performance has been achieved by Villalobos-Castaldi et al [51], where matched filter in a conjugation with entropy-based adaptive thresholding algorithm was employed. The methodology was applied on DRIVE dataset where it used matched filter in sake of piecewise linear segments enhancement of the retina vascular structure.…”

“…Compared to the performance achieved in [51], Chanwimaluang and Fan [49] followed same procedure that was proposed in [51] to extract both the retinal vessel and the optic disk using STARE dataset. However, the time consumed approximated 2.5 min per retina image; most of it was consumed in matched filtering and local entropy thresholding steps.…”

“…However, the time consumed approximated 2.5 min per retina image; most of it was consumed in matched filtering and local entropy thresholding steps. Moreover, it required post-processing steps that were not required by [51] including long filtering stages for isolated pixel removal. Then, filtering steps were followed by morphological thinning used to identify the retinal vascular intersection/crossovers.…”

“…on the other hand, the optic disk identification proceeded into two major stages: (1) optic disk center identification through maximum local variance detection; and (2) optic disk boundary identification through snake active contour. Even though the same methodology steps have been followed by both [49,51], they are extremely different in terms of achieved performance.…”

Survey on Ocular Blood Vessel Segmentation

“…[3] applied reliable vessel extraction that's a prerequisite for subsequent retinal picture evaluation and processing because vessels are the foremost and maximum stable structures performing in those pictures. Accurate segmentation of retinal snap shots influences immediately the performance of trivia extraction.…”