光学学报, 2020, 40 (12): 1210001, 网络出版: 2020-06-03
改进U型网络的眼底视网膜血管分割方法 下载: 2059次
AnImproved Method for Retinal Vascular Segmentation in U-Net
图像处理 视网膜血管 U-Net 循环残差网络 注意力机制 image processing retinal vessels U-Net recurrent residual network attention mechanism
摘要
当前主流的眼底视网膜血管分割方法存在细微血管细粒度特征很难采集和细节容易丢失的问题。为解决这一问题,设计了一种改进U-Net模型算法,该算法将U-Net上下采样中的原始卷积层改为二次循环残差卷积层,提升了特征的使用效率;在解码部分引入多通道注意力模型,改善了低对比度下细小血管的分割效果。该算法在DRIVE (Digital Retinal Images for Vessel Extraction)和STARE (Structured Analysis of the Retina)两个数据库的准确率分别为96.89%和97.96%,敏感度分别为80.28%和82.27%,AUC(Area Under Curve)性能分别为98.41%和98.65%,较现有的先进算法有一定的提升。本文所提算法能有效提高眼底图像细微血管分割准确率。
Abstract
The fine-grained characteristics of blood vessels are difficult to obtain, and the details of the blood vessels are obscured when the current mainstream methods of retinal vascular segmentation are employed. This paper proposes an improved U-Net model algorithm to address these problems. The convolution layer of quadratic-cycle residual difference was used to replace the original convolutional layer in the upper and lower sampling of U-Net to improve the utilization rate of the features. A multichannel attention model was introduced in the decoding part to improve the segmentation effect of small blood vessels with low contrast. Results show that the accuracies of the algorithm in DRIVE (Digital Retinal Images for Vessel Extraction) and STARE (Structured Analysis of the Retina) databases are 96.89% and 97.96%, the sensitivities are 80.28% and 82.27%, and the AUC performances are 98.41% and 98.65%, respectively. All these parameters are higher than those of existing advanced algorithms. The proposed algorithm can effectively improve the segmentation accuracy of fine blood vessels in fundus images.
薛文渲, 刘建霞, 刘然, 袁晓辉. 改进U型网络的眼底视网膜血管分割方法[J]. 光学学报, 2020, 40(12): 1210001. Wenxuan Xue, Jianxia Liu, Ran Liu, Xiaohui Yuan. AnImproved Method for Retinal Vascular Segmentation in U-Net[J]. Acta Optica Sinica, 2020, 40(12): 1210001.