Search

Purpose. Retinal blood vessel image segmentation is an important step in ophthalmological analysis. However, it is difficult to segment small vessels accurately because of low contrast and complex feature information of blood vessels. The objective of this study is to develop an improved retinal blood vessel segmentation structure (WA-Net) to overcome these challenges. Methods. This paper mainly focuses on the width of deep learning. The channels of the ResNet block were broadened to propagate more low-level features, and the identity mapping pathway was slimmed to maintain parameter complexity. A residual atrous spatial pyramid module was used to capture the retinal vessels at various scales. We applied weight normalization to eliminate the impacts of the mini-batch and improve segmentation accuracy. The experiments were performed on the DRIVE and STARE datasets. To show the generalizability of WA-Net, we performed cross-training between datasets. Results. The global accuracy and specificity within datasets were 95.66% and 96.45% and 98.13% and 98.71%, respectively. The accuracy and area under the curve of the interdataset diverged only by 1%∼2% compared with the performance of the corresponding intradataset. Conclusion. All the results show that WA-Net extracts more detailed blood vessels and shows superior performance on retinal blood vessel segmentation tasks. [ABSTRACT FROM AUTHOR]

The analysis of the vascular tree is a fundamental part of the clinical assessment of retinal images. The diversity of blood vessel calibers and curvatures, as well as the ocular vascular alterations derived from the progress of certain diseases, make automatic blood vessel segmentation a challenging task. In this paper, a novel Width Attention-based Convolutional Neural Network, called WA-Net, is proposed. In the WA-Net, a fundus image is decomposed into multiple channels by a layer of Distorted Second-Order Differential Gaussian Matched Filters (DSD-GMF), where each channel is associated with a blood-vessel width. Subsequently, the channel relevance is weighted through the Width Attention Module (WAM), which considers channel and position correlations. Finally, in order to specialize the feature maps with a concrete vessel-width category, either thin-vessel or thick-vessel related, the weighted channels are divided into two groups by the Two-Stream Block, composed of three-level UNet streams. Experimental results on three public datasets (DRIVE/STARE/CHASE) indicate that the proposed method provides a performance gain over other attention-based and non-attention-based architectures, achieving state-of-the-art Accuracy and Area-Under-the-Receiver-Operating-Characteristic-Curve scores (0.9575/0.9665/9653 and 0.9784/0.9865/0.9841 within the Field-of-View, respectively). [Display omitted] • A Width-Aware model improves the state-of-the-art for Retinal Vessel Segmentation. • WA-Net handles blood vessel width variety through a Gaussian-Matched-Filter layer. • WA-Net specializes feature extraction through using a Two-Stream scheme. • WA-Net is more accurate and 13x smaller than other state-of-the-art models. [ABSTRACT FROM AUTHOR]