Skip to main content
QUICK REVIEW

[论文解读] On the Benefit of Dual-domain Denoising in a Self-supervised Low-dose CT Setting

Fabian Wagner, Mareike Thies|arXiv (Cornell University)|Nov 2, 2022
Medical Imaging Techniques and Applications被引用 4
一句话总结

本文提出了一种端到端、自监督的双域去噪流水线,用于低剂量CT,通过在不依赖高剂量真实值的情况下,联合优化投影域和图像域的去噪。在腹部CT上,其PSNR提升最高达94.1%,在XRM数据上相比低剂量基线提升2.9%。结果表明,投影空间中的早期噪声抑制显著提升了最终图像质量。

ABSTRACT

Computed tomography (CT) is routinely used for three-dimensional non-invasive imaging. Numerous data-driven image denoising algorithms were proposed to restore image quality in low-dose acquisitions. However, considerably less research investigates methods already intervening in the raw detector data due to limited access to suitable projection data or correct reconstruction algorithms. In this work, we present an end-to-end trainable CT reconstruction pipeline that contains denoising operators in both the projection and the image domain and that are optimized simultaneously without requiring ground-truth high-dose CT data. Our experiments demonstrate that including an additional projection denoising operator improved the overall denoising performance by 82.4-94.1%/12.5-41.7% (PSNR/SSIM) on abdomen CT and 1.5-2.9%/0.4-0.5% (PSNR/SSIM) on XRM data relative to the low-dose baseline. We make our entire helical CT reconstruction framework publicly available that contains a raw projection rebinning step to render helical projection data suitable for differentiable fan-beam reconstruction operators and end-to-end learning.

研究动机与目标

  • 为解决现有自监督CT去噪方法仅在图像域运行、未能实现投影数据中早期噪声抑制的局限性。
  • 开发一种端到端可训练的流水线,无需配对的高剂量数据,即可联合优化投影域和图像域的去噪。
  • 证明双域去噪相比单域后处理去噪能显著提升图像质量。
  • 发布一个完全可微、开源的螺旋CT数据重排与重建框架,以促进未来对双域及自监督CT流水线的研究。

提出的方法

  • 该方法使用一个可微重建算子R,以及两个可训练的去噪模块:投影域中的D_proj^w和图像域中的D_img^w。
  • 流水线通过反向传播端到端训练,整个链路为:D_proj^w → R → D_img^w → 损失,实现两个去噪器的联合优化。
  • 通过计算从分割投影堆栈独立重建的两个视图之间的自监督损失,实现无需高剂量真实值的训练。
  • 框架包含一个原始投影重排步骤,将螺旋CT数据转换为适用于可微重建的扇形束格式。
  • 该方法支持在两个域中使用任意可微去噪架构(如U-Net、BFS),实现灵活集成。
  • 整个流水线通过基于两个独立投影分割所生成的去噪重建之间特征相似性的对比损失进行训练。
Fig. 1 : Overview of the proposed trainable dual-domain and self-supervised CT denoising pipeline.
Fig. 1 : Overview of the proposed trainable dual-domain and self-supervised CT denoising pipeline.

实验结果

研究问题

  • RQ1与仅在图像域进行去噪相比,将去噪引入投影域是否能提升低剂量CT重建性能?
  • RQ2能否构建一个无需高剂量真实值的自监督、端到端可训练流水线,联合优化投影域和图像域的去噪?
  • RQ3在医学和临床前CT数据上,双域去噪与单域后处理去噪在定量和定性上相比如何?
  • RQ4在投影域中早期抑制噪声在多大程度上简化了后续图像去噪任务?
  • RQ5一个可微、开源的螺旋CT重排与重建框架,是否能促进双域、自监督CT方法的广泛应用?

主要发现

  • 与低剂量基线相比,双域去噪在腹部CT数据上使PSNR提升82.4–94.1%,SSIM提升12.5–41.7%。
  • 在X射线显微断层扫描(XRM)数据上,该方法相比低剂量基线实现1.5–2.9%的PSNR增益和0.4–0.5%的SSIM提升。
  • 双域自监督流水线优于所有单域后处理基线,证实投影域中的早期去噪具有显著优势。
  • 所提出的框架性能接近监督基线,其中双U-Net自监督模型在XRM数据上达到32.5 dB的PSNR。
  • 视觉评估表明,双域模型比仅图像去噪更有效地降低噪声,并更好地保留了解剖细节,尤其在肝脏病灶等低对比度区域表现更优。
  • 开源的螺旋CT重排与可微重建框架支持对原始投影数据进行端到端学习,为未来双域及自监督CT研究提供了有力支持。
Fig. 2 : Illustration of the proposed end-to-end self-supervised denoising pipeline. Projections are split in two stacks and reconstructed separately. The loss calculated between denoised prediction and reconstructed target stack is backpropagated through $\mathbf{D}_{\text{img}}^{w}$ and $\mathbf{R
Fig. 2 : Illustration of the proposed end-to-end self-supervised denoising pipeline. Projections are split in two stacks and reconstructed separately. The loss calculated between denoised prediction and reconstructed target stack is backpropagated through $\mathbf{D}_{\text{img}}^{w}$ and $\mathbf{R

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。