Performance comparison of publicly available retinal blood vessel segmentation methods

Abstract

Retinal blood vessel structure is an important indicator of many retinal and systemic diseases, which has motivated the development of various image segmentation methods for the blood vessels. In this study, two supervised and three unsupervised segmentation methods with a publicly available implementation are reviewed and quantitatively compared with each other on five public databases with ground truth segmentation of the vessels. Each method is tested under consistent conditions with two types of preprocessing, and the parameters of the methods are optimized for each database. Additionally, possibility to predict the parameters of the methods by the linear regression model is tested for each database. Resolution of the input images and amount of the vessel pixels in the ground truth are used as predictors. The results show the positive influence of preprocessing on the performance of the unsupervised methods. The methods show similar performance for segmentation accuracy, with the best performance achieved by the method by Azzopardi et al. (Acc 94.0) on ARIADB, the method by Soares et al. (Acc 94.6, 94.7) on CHASEDB1 and DRIVE, and the method by Nguyen et al. (Acc 95.8, 95.5) on HRF and STARE. The method by Soares et al. performed better with regard to the area under the ROC curve. Qualitative differences between the methods are discussed. Finally, it was possible to predict the parameter settings that give performance close to the optimized performance of each method.

A proposal to measure the similarity between retinal vessel segmentations images

“…These are the so-called true DRIVE and STARE which are currently the most common datasets for the evaluation of retinal vessel segmentation methods, can be found in Refs. (Fraz et al, 2012;Garg & Gupta, 2016;Vostatek et al, 2017).…”

“…Finally, we highlight that our approach is compatible, and can be used in a complementary manner, with similarity assessment approaches which are based on other aspects the image's structure such as connectivity, area, and the length of the segmented vessels (Vostatek et al, 2017), or the so-called skeleton maps (Fraz et al, 2012;Kirbas, C. Quek, 2004).…”

“…A particularly useful reference for the present work is Ref. (Vostatek et al, 2017) (Garg & Gupta, 2016). Other more specific categorizations propose vessel tracking, multi-scale approaches, and vessel profiling as independent classifications from the more general morphology-based techniques (Fraz et al, 2012;Kirbas, C. Quek, 2004), and intensity-based segmentation algorithm (Gonzalez & Woods, 2001).…”

Recent Development on Detection Methods for the Diagnosis of Diabetic Retinopathy

“…A sensitivity (SE) of 96.1% and specificity (SP) of 92.2% in the classification of true vessel points was reported. A comprehensive review of supervised-related, unsupervised-related, image processing-related, and data miningrelated algorithms is presented in [45][46][47] with high vessel segmentation results. The study in [46] combined normal and abnormal retinal features with the patient's contextual information at adaptive granularity levels through data mining methods to segment retinal vessels.…”

State of the Art in Vascular Biometrics

“…As fundus imaging is used as an important diagnostic tool in (human) medicine (see Sect. 1.8), where the vascular network is mainly targeted as the entity diagnosis is based on, a significant corpus of medical literature exists on techniques to reliably extract the vessel structure (see [260] for a performance comparison of publicly available retinal blood vessel segmentation methods). A wide variety of techniques has been developed, e.g.…”