论文标题
使用合成训练数据集的临床CT图像的微型CT图像辅助交叉模态超分辨率
Micro CT Image-Assisted Cross Modality Super-Resolution of Clinical CT Images Utilizing Synthesized Training Dataset
论文作者
论文摘要
本文提出了一种新颖的，无监督的超分辨率（SR）方法，用于将临床CT的SR进行到微CT（$μ$ CT）的分辨率水平。肺癌的精确非侵入性诊断通常利用临床CT数据。由于临床CT的分辨率限制（大约0.5美元\倍0.5倍0.5 $ mm $ $^3 $），因此很难获得足够的病理信息，例如ALVEOLI级别的入侵区域。另一方面，$μ$ CT扫描允许以更高分辨率（$ 50 \ times 50 \ times50μ{\ rm m}^3 $或更高的分辨率获得肺部标本量（$ 50 \ times 50 \ times）。因此，临床CT体积的超分辨率可能有助于诊断肺癌。典型的SR方法需要对齐对训练的低分辨率（LR）和高分辨率（HR）图像。不幸的是，获得人肺组织的配对临床CT和$ $ $ CT的体积是不可行的。不需要配对的LR和HR图像需要无监督的SR方法。在本文中，我们通过通过修改Cyclegan模拟$μ$ CT图像中的临床CT图像来创建相应的临床CT- $ $ CT对。之后，我们使用模拟的临床CT- $ $ CT图像对来训练基于SRGAN的SR网络。最后，我们使用训练有素的SR网络执行临床CT图像的SR。我们将提出的方法与另一种无监督的SR方法进行比较，用于临床CT图像，称为SR-Cyclegan。实验结果表明，所提出的方法可以成功执行$ $ $ CT水平分辨率的肺癌患者的临床CT图像，并定量和定性地表现优于常规方法（SR-Cyclegan），从而改善了SSIM（结构相似性）0.40至0.51。
This paper proposes a novel, unsupervised super-resolution (SR) approach for performing the SR of a clinical CT into the resolution level of a micro CT ($μ$CT). The precise non-invasive diagnosis of lung cancer typically utilizes clinical CT data. Due to the resolution limitations of clinical CT (about $0.5 \times 0.5 \times 0.5$ mm$^3$), it is difficult to obtain enough pathological information such as the invasion area at alveoli level. On the other hand, $μ$CT scanning allows the acquisition of volumes of lung specimens with much higher resolution ($50 \times 50 \times 50 μ{\rm m}^3$ or higher). Thus, super-resolution of clinical CT volume may be helpful for diagnosis of lung cancer. Typical SR methods require aligned pairs of low-resolution (LR) and high-resolution (HR) images for training. Unfortunately, obtaining paired clinical CT and $μ$CT volumes of human lung tissues is infeasible. Unsupervised SR methods are required that do not need paired LR and HR images. In this paper, we create corresponding clinical CT-$μ$CT pairs by simulating clinical CT images from $μ$CT images by modified CycleGAN. After this, we use simulated clinical CT-$μ$CT image pairs to train an SR network based on SRGAN. Finally, we use the trained SR network to perform SR of the clinical CT images. We compare our proposed method with another unsupervised SR method for clinical CT images named SR-CycleGAN. Experimental results demonstrate that the proposed method can successfully perform SR of clinical CT images of lung cancer patients with $μ$CT level resolution, and quantitatively and qualitatively outperformed conventional method (SR-CycleGAN), improving the SSIM (structure similarity) form 0.40 to 0.51.
