Comparative study of retinal vessel segmentation methods on a new publicly available database

By Meindert Niemeijer, Joes Staal, B. van Ginneken, Marco Loog, Michael David Abràmoff. SPIE Medical Imaging, 2004

906 citations recorded by the source corpus.

In this work we compare the performance of a number of vessel segmentation algorithms on a newly constructed retinal vessel image database. Retinal vessel segmentation is important for the detection of numerous eye diseases and plays an important role in automatic retinal disease screening systems. A large number of methods for retinal vessel segmentation have been published, yet an evaluation of these methods on a common database of screening images has not been performed. To compare the performance of retinal vessel segmentation methods we have constructed a large database of retinal images. The database contains forty images in which the vessel trees have been manually segmented. For twenty of those forty images a second independent manual segmentation is available. This allows for a comparison between the performance of automatic methods and the performance of a human observer. The database is available to the research community. Interested researchers are encouraged to upload their segmentation results to our website (http://www.isi.uu.nl/Research/Databases). The performance of five different algorithms has been compared. Four of these methods have been implemented as described in the literature. The fifth pixel classification based method was developed specifically for the segmentation of retinal vessels and is the only supervised method in this test. We define the segmentation accuracy with respect to our gold standard as the performance measure. Results show that the pixel classification method performs best, but the second observer still performs significantly better.

Subfields: Computer Vision and Pattern Recognition, Ophthalmology, Radiology, Nuclear Medicine and Imaging

View Paper: https://www.semanticscholar.org/paper/798e312dd67798024da74f9a8f92946af88c7cd4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Ophthalmology

In Radiology, Nuclear Medicine and Imaging

Citations in the corpus

Retinal Image Analysis Using Curvelet Transform and Multistructure Elements Morphology by Reconstruction (2011, methodology): “In fact, vessels clearly stand out from background in green channel for images of DRIVE database [19].”
An Ensemble Classification-Based Approach Applied to Retinal Blood Vessel Segmentation (2012, methodology): “Niemeijer et al. [20] extracted a feature vector for each pixel that consists of the Gaussian and its derivatives at multiple scales, augmented with the green plane of the RGB image, and then uses a K-nearest neighbor algorithm to estimate the probability of the pixel belonging to a vessel.”
A Discriminatively Trained Fully Connected Conditional Random Field Model for Blood Vessel Segmentation in Fundus Images (2017, neutral): “Several classiﬁers have been considered in the literature, including k -nearest neighbors [6], Bayesian [7], support vector machines (SVMs) [8], [9], neural networks [10], [11], decision trees [12], [13], Gaussian mixture models [5], AdaBoost [14], among others.”
An improved matched filter for blood vessel detection of digital retinal images (2007, methodology): “Many methods, however, have been proposed for this purpose; matched ﬁlter [1], morphology edge detection [6], ridge-based vessel segmentation [7] which is called pixel classiﬁcation in [8], and wavelet transform [9].”
Retinal Blood Vessel Segmentation Using Line Operators and Support Vector Classification (2007, methodology): “[15] with the aim of comparing the performance of different seg-”
Blood Vessel Segmentation of Fundus Images by Major Vessel Extraction and Subimage Classification (2015, methodology): “In [24], the k-nearest neighbor (kNN) classifier uses a 31-feature set extracted using Gaussians and their derivatives.”
Fast Retinal Vessel Detection and Measurement Using Wavelets and Edge Location Refinement (2012, methodology): “To begin, a ‘vesselness map’ generated by filtering the image is thresholded to give a binary image [14], which is cleaned up and thinned to provide vessel centrelines.”
Retinal Vessel Centerline Extraction Using Multiscale Matched Filters, Confidence and Edge Measures (2006, methodology): “In addition, matched-filterbased LRV is significantly better (p 0.0001 for STARE and p 0.005 for DRIVE) than the Frangi-based LRV.”
FABC: Retinal Vessel Segmentation Using AdaBoost (2010, methodology): “Niemeijer et al. [4] extract a simple feature vector for each pixel from the green plane, and then, use a k-nearest neighbor (kNN) algorithm to estimate the probability of the pixel belonging to a vessel.”
DUNet: A deformable network for retinal vessel segmentation (2018, background): “Niemeijer et al. compared a number of vessel segmentation algorithms [10].”