Glaucoma is one of the leading causes of irreversible blindness. Segmentation of optic disc (OD) and optic cup (OC) on fundus images is a crucial step in glaucoma screening. Although many deep learning models have been constructed for this task, it remains challenging to train an OD/OC segmentation model that could be deployed successfully to different healthcare centers. The difficulties mainly comes from the domain shift issue, i.e., the fundus images collected at these centers usually vary greatly in the tone, contrast, and brightness. To address this issue, in this paper, we propose a novel unsupervised domain adaptation (UDA) method called Reconstruction-driven Dynamic Refinement Network (RDR-Net), where we employ a due-path segmentation backbone for simultaneous edge detection and region prediction and design three modules to alleviate the domain gap. The reconstruction alignment (RA) module uses a variational auto-encoder (VAE) to reconstruct the input image and thus boosts the image representation ability of the network in a self-supervised way. It also uses a style-consistency constraint to force the network to retain more domain-invariant information. The low-level feature refinement (LFR) module employs input-specific dynamic convolutions to suppress the domain-variant information in the obtained low-level features. The prediction-map alignment (PMA) module elaborates the entropy-driven adversarial learning to encourage the network to generate source-like boundaries and regions. We evaluated our RDR-Net against state-of-the-art solutions on four public fundus image datasets. Our results indicate that RDR-Net is superior to competing models in both segmentation performance and generalization ability
翻译:青光眼是导致不可逆失明的主要原因之一。在青光眼筛查中,眼底图像的视盘(OD)和视杯(OC)的分割是一个至关重要的步骤。尽管已经构建了许多深度学习模型用于此任务,但训练一个可在不同医疗中心成功部署的 OD/OC 分割模型仍然具有挑战性。这主要是因为领域偏移问题,即这些中心收集的眼底图像通常在色调、对比度和亮度方面变化很大。为了解决这个问题,本文提出了一种名为“重构驱动的动态精细化网络”(RDR-Net)的新型无监督域自适应(UDA)方法,在分割骨干网络中使用两条路径,同时进行边缘检测和区域预测,设计了三个模块来减轻领域差距。重构对齐(RA)模块使用变分自编码器(VAE)重构输入图像,以自我监督的方式提升网络的图像表示能力。它还使用样式一致性约束来强制网络保留更多的领域不变信息。低层特征精细化(LFR)模块采用输入特定的动态卷积来抑制获得的低层特征中的领域特定信息。预测图映射对齐(PMA)模块采用熵驱动的对抗学习技术,鼓励网络生成类似源域的边界和区域。我们在四个公共眼底图像数据集上评估了我们的 RDR-Net,并与现有的解决方案进行了比较。实验结果表明,RDR-Net 在分割性能和泛化能力上优于其他模型。