Blood vessel classification into arteries and veins in retinal images

2007 ◽  
Author(s):  
Claudia Kondermann ◽  
Daniel Kondermann ◽  
Michelle Yan
Keyword(s):  
Blood Vessel ◽  
Retinal Images ◽  
Arteries And Veins

A Comparative Study of Different Blood Vessel Detection on Retinal Images

2019 ◽  
Vol 11 (4) ◽  
Author(s):  
Wan Azani Mustafa ◽  
◽  
Ahmad Syauqi Mahmud ◽  
Muhammad Zaid Aihsan ◽  
M. Saifizi ◽  
...  
Keyword(s):  
Blood Vessel ◽  
Comparative Study ◽  
Retinal Images ◽  
Blood Vessel Detection ◽  
Vessel Detection

scholarly journals Extraction of Blood Vessels in Retinal Images Using Four Different Techniques

2013 ◽  
Vol 2013 ◽  
pp. 1-21 ◽  
Author(s):  
Asloob Ahmad Mudassar ◽  
Saira Butt
Keyword(s):  
Blood Vessel ◽  
Cross Sections ◽  
Retinal Images ◽  
Matched Filtering ◽  
Edge Enhancement ◽  
Promising Technique ◽  
Filtering Technique ◽  
Continuation Algorithm ◽  
Line Cross ◽  
Vessel Extraction

A variety of blood vessel extraction (BVE) techniques exist in the literature, but they do not always lead to acceptable solutions especially in the presence of anomalies where the reported work is limited. Four techniques are presented for BVE: (1) BVE using Image Line Cross-Sections (ILCS), (2) BVE using Edge Enhancement and Edge Detection (EEED), (3) BVE using Modified Matched Filtering (MMF), and (4) BVE using Continuation Algorithm (CA). These four techniques have been designed especially for abnormal retinal images containing low vessel contrasts, drusen, exudates, and other artifacts. The four techniques were applied to 30 abnormal retinal images, and the success rate was found to be (95 to 99%) for CA, (88–91%) for EEED, (80–85%) for MMF, and (74–78%) for ILCS. Application of these four techniques to 105 normal retinal images gave improved results: (99-100%) for CA, (96–98%) for EEED, (94-95%) for MMF, and (88–93%) for ILCS. Investigations revealed that the four techniques in the order of increasing performance could be arranged as ILCS, MMF, EEED, and CA. Here we demonstrate these four techniques for abnormal retinal images only. ILCS, EEED, and CA are novel additions whereas MMF is an improved and modified version of an existing matched filtering technique. CA is a promising technique.

scholarly journals Reliability of Manual Vascular Segmentation for Retinal Fractal Dimension using Peek Retinatm

2019 ◽  
Vol 8 (9S3) ◽  
pp. 1560-1564
Keyword(s):  
Fractal Dimension ◽  
Blood Vessel ◽  
Blood Vessels ◽  
Retinal Image ◽  
Retinal Images ◽  
Retinal Vasculature ◽  
Manual Segmentation ◽  
Good Reliability ◽  
Fundus Camera ◽  
Df Analysis

Fractal dimension (Df) has been identified as indirect measure in quantifying the complexity of retinal vessel network which is useful for early detection of vascular changes. Reliability studies of Df measurement on retinal vasculature, has been conducted on retinal images processed by using semi-automated software which only permits image with 45ᵒ field of view (FOV). Smartphone-assisted fundus camera retinal image has a maximum 30ᵒ FOV which warrant manual processing in measuring the Df. Retinal blood vessels need to be manually segmented to produce binary images for retinal vasculatures Df measurement. Therefore, this study was conducted to determine the intragrader and intergrader reliability of manual segmentation of the retinal vasculature Df measurement from retinal images taken using a smartphone-assisted fundus camera Forty-five retinal images were captured using the Portable Eye Examination Kit Retina (Peek Retina™, Peek Vision Ltd, UK). Suitable image for Df analysis were selected based on gradable retinal image criteria which included; i) good image focus, ii) centered position of optic nerve head (ONH) and iii) significant blood vessel visibility. The images were cropped 0.5 disc diameters away from disc margin and resized to 500x500 pixels using GNU Image Manipulation Program Version 2.8.18 (GIMP, The GIMP Team, United States). Retinal vessels were manually traced by using layering capabilities for blood vessel segmentation. Df values of segmented blood vessels were measured by using Image J (National Institutes of Health, USA) and its plugin software, FracLac Version 2.5. Intragrader and intergrader reliability was determined by comparing the Df values between; two readings measured one week apart by a grader and readings from two different graders, respectively, using intraclass correlation coefficient (ICC) and Bland-Altman graphical plots. Intragrader agreement for retinal Df showed good reliability with ICC of 0.899 (95% CI: 0.814–0.945). Bland Altman analysis indicated good agreement between Df values at different grading time (mean difference 0.0050; 95% CI:-0.0001–0.0101). Intergrader reliability for retinal Df was high with ICC of 0.814 (95% CI: 0.459–0.919). Bland Altman plot revealed good intergrader agreement for retinal Df between two graders with a bias value of 0.0158 (95% CI: 0.0092–0.0223). In conclusion, manual segmentation of retinal image captured by smartphone-assisted fundus camera has good reliability (0.75 < ICC < 0.9) for Df analysis to study the morphology of retinal vasculatures.

scholarly journals Semi-Supervised Learning Method of U-Net Deep Learning Network for Blood Vessel Segmentation in Retinal Images

Symmetry ◽  
2020 ◽  
Vol 12 (7) ◽  
pp. 1067
Author(s):  
Dali Chen ◽  
Yingying Ao ◽  
Shixin Liu
Keyword(s):  
Deep Learning ◽  
Blood Vessel ◽  
Supervised Learning ◽  
Vessel Segmentation ◽  
Training Dataset ◽  
Retinal Images ◽  
Learning Method ◽  
Learning Network ◽  
Blood Vessel Segmentation ◽  
Deep Learning Network

Blood vessel segmentation methods based on deep neural networks have achieved satisfactory results. However, these methods are usually supervised learning methods, which require large numbers of retinal images with high quality pixel-level ground-truth labels. In practice, the task of labeling these retinal images is very costly, financially and in human effort. To deal with these problems, we propose a semi-supervised learning method which can be used in blood vessel segmentation with limited labeled data. In this method, we use the improved U-Net deep learning network to segment the blood vessel tree. On this basis, we implement the U-Net network-based training dataset updating strategy. A large number of experiments are presented to analyze the segmentation performance of the proposed semi-supervised learning method. The experiment results demonstrate that the proposed methodology is able to avoid the problems of insufficient hand-labels, and achieve satisfactory performance.

Sign in / Sign up

Export Citation Format

Share Document