A new hybrid algorithm for retinal vessels segmentation on fundus images

Automatic retinal vessels segmentation is an important task in medical applications. However, most of the available retinal vessels segmentation methods are prone to poorer results when dealing with challenging situations such as detecting low-contrast micro-vessels, vessels with central reflex, and...

Full description

Bibliographic Details
Main Authors: Dharmawan, Dhimas Arief, Li, Di, Ng, Boon Poh, Rahardja, Susanto
Other Authors: School of Electrical and Electronic Engineering
Format: Journal Article
Language:English
Published: 2019
Subjects:
Online Access:https://hdl.handle.net/10356/106276
http://hdl.handle.net/10220/48958
http://dx.doi.org/10.1109/ACCESS.2019.2906344
_version_ 1824455215135326208
author Dharmawan, Dhimas Arief
Li, Di
Ng, Boon Poh
Rahardja, Susanto
author2 School of Electrical and Electronic Engineering
author_facet School of Electrical and Electronic Engineering
Dharmawan, Dhimas Arief
Li, Di
Ng, Boon Poh
Rahardja, Susanto
author_sort Dharmawan, Dhimas Arief
collection NTU
description Automatic retinal vessels segmentation is an important task in medical applications. However, most of the available retinal vessels segmentation methods are prone to poorer results when dealing with challenging situations such as detecting low-contrast micro-vessels, vessels with central reflex, and vessels in the presence of pathologies. This paper presents a new hybrid algorithm for retinal vessels segmentation on fundus images. The proposed algorithm overcomes the difficulty when dealing with the challenging situations by first applying a new directionally sensitive blood vessel enhancement method before sending fundus images to a convolutional neural network architecture derived from U-Net. To train and test the algorithm, fundus images from the DRIVE and STARE databases, as well as high-resolution fundus images from the HRF database, are utilized. In the experiment, the proposed algorithm outperforms the state-of-the-art methods in four major measures, i.e., sensitivity, F1-score, G-mean, and Mathews correlation coefficient both on the low- and high-resolution images. In addition, the proposed algorithm achieves the best connectivity-area-length score among the competing methods. Given such performance, the proposed algorithm can be adapted for vessel-like structures segmentation in other medical applications. In addition, since the new blood vessel enhancement method is independent of the U-Net model, it can be easily applied to other deep learning architectures.
first_indexed 2025-02-19T03:34:40Z
format Journal Article
id ntu-10356/106276
institution Nanyang Technological University
language English
last_indexed 2025-02-19T03:34:40Z
publishDate 2019
record_format dspace
spelling ntu-10356/1062762019-12-06T22:07:52Z A new hybrid algorithm for retinal vessels segmentation on fundus images Dharmawan, Dhimas Arief Li, Di Ng, Boon Poh Rahardja, Susanto School of Electrical and Electronic Engineering CNN DRNTU::Engineering::Electrical and electronic engineering Fundus Images Automatic retinal vessels segmentation is an important task in medical applications. However, most of the available retinal vessels segmentation methods are prone to poorer results when dealing with challenging situations such as detecting low-contrast micro-vessels, vessels with central reflex, and vessels in the presence of pathologies. This paper presents a new hybrid algorithm for retinal vessels segmentation on fundus images. The proposed algorithm overcomes the difficulty when dealing with the challenging situations by first applying a new directionally sensitive blood vessel enhancement method before sending fundus images to a convolutional neural network architecture derived from U-Net. To train and test the algorithm, fundus images from the DRIVE and STARE databases, as well as high-resolution fundus images from the HRF database, are utilized. In the experiment, the proposed algorithm outperforms the state-of-the-art methods in four major measures, i.e., sensitivity, F1-score, G-mean, and Mathews correlation coefficient both on the low- and high-resolution images. In addition, the proposed algorithm achieves the best connectivity-area-length score among the competing methods. Given such performance, the proposed algorithm can be adapted for vessel-like structures segmentation in other medical applications. In addition, since the new blood vessel enhancement method is independent of the U-Net model, it can be easily applied to other deep learning architectures. Published version 2019-06-26T07:05:44Z 2019-12-06T22:07:52Z 2019-06-26T07:05:44Z 2019-12-06T22:07:52Z 2019 Journal Article Dharmawan, D. A., Li, D., Ng, B. P., & Rahardja, S. (2019). A new hybrid algorithm for retinal vessels segmentation on fundus images. IEEE Access, 7, 41885-41896. doi: 10.1109/ACCESS.2019.2906344 https://hdl.handle.net/10356/106276 http://hdl.handle.net/10220/48958 http://dx.doi.org/10.1109/ACCESS.2019.2906344 en IEEE Access © 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information. 12 p. application/pdf
spellingShingle CNN
DRNTU::Engineering::Electrical and electronic engineering
Fundus Images
Dharmawan, Dhimas Arief
Li, Di
Ng, Boon Poh
Rahardja, Susanto
A new hybrid algorithm for retinal vessels segmentation on fundus images
title A new hybrid algorithm for retinal vessels segmentation on fundus images
title_full A new hybrid algorithm for retinal vessels segmentation on fundus images
title_fullStr A new hybrid algorithm for retinal vessels segmentation on fundus images
title_full_unstemmed A new hybrid algorithm for retinal vessels segmentation on fundus images
title_short A new hybrid algorithm for retinal vessels segmentation on fundus images
title_sort new hybrid algorithm for retinal vessels segmentation on fundus images
topic CNN
DRNTU::Engineering::Electrical and electronic engineering
Fundus Images
url https://hdl.handle.net/10356/106276
http://hdl.handle.net/10220/48958
http://dx.doi.org/10.1109/ACCESS.2019.2906344
work_keys_str_mv AT dharmawandhimasarief anewhybridalgorithmforretinalvesselssegmentationonfundusimages
AT lidi anewhybridalgorithmforretinalvesselssegmentationonfundusimages
AT ngboonpoh anewhybridalgorithmforretinalvesselssegmentationonfundusimages
AT rahardjasusanto anewhybridalgorithmforretinalvesselssegmentationonfundusimages
AT dharmawandhimasarief newhybridalgorithmforretinalvesselssegmentationonfundusimages
AT lidi newhybridalgorithmforretinalvesselssegmentationonfundusimages
AT ngboonpoh newhybridalgorithmforretinalvesselssegmentationonfundusimages
AT rahardjasusanto newhybridalgorithmforretinalvesselssegmentationonfundusimages