[论文解读] DU-GAN: Generative Adversarial Networks with Dual-Domain U-Net Based Discriminators for Low-Dose CT Denoising

Zhizhong Huang, Junping Zhang+2 more|arXiv (Cornell University)|Aug 24, 2021

Medical Imaging Techniques and Applications参考文献 45被引用 199

一句话总结

本文提出 DU-GAN，一种基于双域 U-Net 判别器的生成对抗网络，用于低剂量 CT 噪声抑制。该方法在图像域和梯度域中均采用 U-Net 判别器，以捕捉全局结构与局部细节，从而提升去噪质量并减少条纹伪影；该方法在模拟数据集和真实世界数据集上均实现了最先进性能，同时通过 CutMix 增强的置信度图实现了边缘保留与不确定性可视化。

ABSTRACT

LDCT has drawn major attention in the medical imaging field due to the potential health risks of CT-associated X-ray radiation to patients. Reducing the radiation dose, however, decreases the quality of the reconstructed images, which consequently compromises the diagnostic performance. Various deep learning techniques have been introduced to improve the image quality of LDCT images through denoising. GANs-based denoising methods usually leverage an additional classification network, i.e. discriminator, to learn the most discriminate difference between the denoised and normal-dose images and, hence, regularize the denoising model accordingly; it often focuses either on the global structure or local details. To better regularize the LDCT denoising model, this paper proposes a novel method, termed DU-GAN, which leverages U-Net based discriminators in the GANs framework to learn both global and local difference between the denoised and normal-dose images in both image and gradient domains. The merit of such a U-Net based discriminator is that it can not only provide the per-pixel feedback to the denoising network through the outputs of the U-Net but also focus on the global structure in a semantic level through the middle layer of the U-Net. In addition to the adversarial training in the image domain, we also apply another U-Net based discriminator in the image gradient domain to alleviate the artifacts caused by photon starvation and enhance the edge of the denoised CT images. Furthermore, the CutMix technique enables the per-pixel outputs of the U-Net based discriminator to provide radiologists with a confidence map to visualize the uncertainty of the denoised results, facilitating the LDCT-based screening and diagnosis. Extensive experiments on the simulated and real-world datasets demonstrate superior performance over recently published methods both qualitatively and quantitatively.

研究动机与目标

• 解决因辐射剂量降低导致低剂量 CT（LDCT）图像中噪声和条纹伪影的问题。
• 通过增强全局结构与局部细节学习，超越传统 GAN 的去噪性能。
• 为放射科医生提供去噪 LDCT 图像中每个像素的置信度图，以实现不确定性可视化。
• 通过梯度域对抗训练减少光子饥饿导致的伪影。
• 通过 CutMix 数据增强提升判别器的鲁棒性与泛化能力。

提出的方法

• 提出一种双域 GAN 框架，结合图像域与梯度域中的 U-Net 判别器，学习去噪图像与标准剂量 CT 图像之间的多尺度差异。
• 采用输出每个像素置信度图的 U-Net 判别器，以提供局部细节反馈，并通过瓶颈层实现全局表征，确保结构一致性。
• 在梯度域中引入第二个 U-Net 判别器，以增强边缘细节并抑制光子饥饿引起的条纹伪影。
• 应用 CutMix 数据增强，混合真实与生成的图像块，提升判别器泛化能力并支持不确定性估计。
• 采用多损失训练目标，结合对抗损失、感知损失与 L1 损失，以平衡保真度、纹理与结构质量。
• 在 64×64 图像块上进行训练以降低计算成本，同时推理阶段在全分辨率下执行。

实验结果

研究问题

• RQ1 与标准 GAN 判别器相比，基于 U-Net 的判别器是否能同时提升 LDCT 去噪中局部细节与全局结构的学习能力？
• RQ2 双域（图像与梯度）对抗训练是否能增强 LDCT 图像的边缘保留并减少伪影？
• RQ3 经过 CutMix 增强的 U-Net 判别器能否提供可靠的像素级不确定性图以供临床解读？
• RQ4 图像块大小如何影响 U-Net 判别器在 LDCT 去噪中的性能表现？
• RQ5 所提出的 DU-GAN 是否能在不同剂量水平与真实世界数据集上实现良好泛化？

主要发现

• 基于 U-Net 的判别器在所有判别器类型中取得了最高的 SSIM 分数（0.7454），优于图像块、像素与全局判别器。
• 采用双域训练的方法在 Mayo-10% 数据集上达到 PSNR 23.1102、RMSE 0.0724 与 SSIM 0.7343，优于所有消融实验变体。
• 使用 CutMix 后，判别器对局部细节的关注度提升，SSIM 增加，PSNR 与 RMSE 略有下降。
• 较小的图像块尺寸（64×64）优于较大的尺寸，PSNR 达 22.3075，SSIM 达 0.7489。
• 双域训练有效减少了条纹伪影并增强了边缘锐度，经视觉与定量分析验证。
• 该方法在不同剂量水平的模拟数据集以及真实世界数据集上均表现出良好泛化能力，且架构与超参数保持一致。