[论文解读] D2A U-Net: Automatic Segmentation of COVID-19 Lesions from CT Slices with Dilated Convolution and Dual Attention Mechanism

Xiangyu Zhao, Peng Zhang+7 more|arXiv (Cornell University)|Feb 10, 2021

COVID-19 diagnosis using AI参考文献 30被引用 6

一句话总结

该论文提出D2A U-Net，一种深度学习模型，结合空洞卷积与双注意力机制（GAM和DAM），以提升CT扫描中COVID-19肺部病灶的自动分割性能。该方法增强了特征表示与感受野，采用预训练的ResNeXt-50主干网络时，Dice分数达到0.7298，显著减少了假阳性并提高了召回率。

ABSTRACT

Coronavirus Disease 2019 (COVID-19) has caused great casualties and becomes almost the most urgent public health events worldwide. Computed tomography (CT) is a significant screening tool for COVID-19 infection, and automated segmentation of lung infection in COVID-19 CT images will greatly assist diagnosis and health care of patients. However, accurate and automatic segmentation of COVID-19 lung infections remains to be challenging. In this paper we propose a dilated dual attention U-Net (D2A U-Net) for COVID-19 lesion segmentation in CT slices based on dilated convolution and a novel dual attention mechanism to address the issues above. We introduce a dilated convolution module in model decoder to achieve large receptive field, which refines decoding process and contributes to segmentation accuracy. Also, we present a dual attention mechanism composed of two attention modules which are inserted to skip connection and model decoder respectively. The dual attention mechanism is utilized to refine feature maps and reduce semantic gap between different levels of the model. The proposed method has been evaluated on open-source dataset and outperforms cutting edges methods in semantic segmentation. Our proposed D2A U-Net with pretrained encoder achieves a Dice score of 0.7298 and recall score of 0.7071. Besides, we also build a simplified D2A U-Net without pretrained encoder to provide a fair comparison with other models trained from scratch, which still outperforms popular U-Net family models with a Dice score of 0.7047 and recall score of 0.6626. Our experiment results have shown that by introducing dilated convolution and dual attention mechanism, the number of false positives is significantly reduced, which improves sensitivity to COVID-19 lesions and subsequently brings significant increase to Dice score.

研究动机与目标

• 为解决CT图像中COVID-19肺部病灶准确且自动分割的挑战，该挑战因病灶形态多变及边界模糊而难以克服。
• 通过增强特征表示与模型注意力，减少假阳性并提高对小病灶或定义不明确病灶的敏感性。
• 开发一种基于U-Net的架构，集成空洞卷积与双注意力模块，以在性能上超越现有最先进模型。
• 在开源数据集上评估模型性能，并与U-Net变体（包括从零开始训练的模型）进行比较。
• 通过消融研究探索各组件（GAM、RAB、预训练主干）的贡献。

提出的方法

• 模型在解码器中采用空洞卷积模块，以扩展感受野并改善不同大小病灶的分割效果。
• 引入一种新型门控注意力模块（GAM），通过融合编码器特征与语义丰富的门控信号，优化跳跃连接。
• 在解码器中应用解码器注意力模块（DAM），以优化上采样后的特征图，增强对边界模糊或微小病灶的表征能力。
• 残差注意力块（RAB）将混合空洞卷积与DAM结合，同时在解码器中扩大感受野并提升特征精炼能力。
• 采用双注意力机制，独立作用于跳跃连接与解码器特征，以减小编码器与解码器层级之间的语义差距。
• 架构采用VGG风格主干网络与更深的ResNeXt-50（32×4d）主干网络进行评估，涵盖有无迁移学习两种情况。

实验结果

研究问题

• RQ1 空洞卷积与双注意力机制的结合是否能提升CT扫描中COVID-19肺部病灶的分割精度？
• RQ2 与标准U-Net跳跃连接相比，所提出的门控注意力模块（GAM）在跳跃连接中如何增强特征表示？
• RQ3 解码器注意力模块（DAM）在病灶边界模糊或不清晰时，其对分割性能的提升程度如何？
• RQ4 结合空洞卷积与注意力的残差注意力块（RAB）是否在病灶分割中优于标准U-Net模块？
• RQ5 通过消融研究，各组件（GAM、RAB、预训练主干）对整体性能的贡献程度如何？

主要发现

• 采用预训练的ResNeXt-50（32×4d）主干网络的D2A U-Net模型，Dice分数达0.7298，召回分数达0.7071，优于现有最先进模型。
• 即使不使用预训练主干网络，简化版D2A U-Net的Dice分数为0.7047，召回率为0.6626，仍优于从零开始训练的主流U-Net变体。
• 消融研究证实，门控注意力模块（GAM）显著提升性能，Dice分数从U-Net的0.6384提升至0.6771。
• GAM与RAB的结合产生协同增益效应，Dice分数提升至0.7047，超过两者单独贡献之和。
• 使用预训练主干网络进一步提升性能，像素误差从0.0323降至0.0311，召回率提升至0.7071。
• 视觉对比显示，与U-Net、Attention U-Net及U-Net++相比，D2A U-Net生成的分割掩码更精确、更敏感，尤其在微小或模糊病灶上表现更优。