Liang Xiaokun, Li Na, Zhang Zhicheng, Yu Shaode, Qin Wenjian, Li Yafen, Chen Shupeng, Zhang Huailing, Xie Yaoqin
Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen 518055, China.
Shenzhen College of Advanced Technology, University of Chinese Academy of Sciences, Shenzhen 518055, China.
Quant Imaging Med Surg. 2019 Jul;9(7):1242-1254. doi: 10.21037/qims.2019.05.19.
Shading artifact may lead to CT number inaccuracy, image contrast loss and spatial non-uniformity (SNU), which is considered as one of the fundamental limitations for volumetric CT (VCT) application. To correct the shading artifact, a novel approach is proposed using deep learning and an adaptive filter (AF).
Firstly, we apply the deep convolutional neural network (DCNN) to train a human tissue segmentation model. The trained model is implemented to segment the tissue. According to the general knowledge that CT number of the same human tissue is approximately the same, a template image without shading artifact can be generated using segmentation and then each tissue is filled with the corresponding CT number of a specific tissue. By subtracting the template image from the uncorrected image, the residual image with image detail and shading artifact are generated. The shading artifact is mainly low-frequency signals while the image details are mainly high-frequency signals. Therefore, we proposed an adaptive filter to separate the shading artifact and image details accurately. Finally, the estimated shading artifacts are deleted from the raw image to generate the corrected image.
On the Catphan©504 study, the error of CT number in the corrected image's region of interest (ROI) is reduced from 109 to 11 HU, and the image contrast is increased by a factor of 1.46 on average. On the patient pelvis study, the error of CT number in selected ROI is reduced from 198 to 10 HU. The SNU calculated from the ROIs decreases from 24% to 9% after correction.
The proposed shading correction method using DCNN and AF may find a useful application in future clinical practice.
阴影伪影可能导致CT数值不准确、图像对比度丧失和空间不均匀性（SNU），这被认为是容积CT（VCT）应用的基本限制之一。为了校正阴影伪影，提出了一种使用深度学习和自适应滤波器（AF）的新方法。
首先，我们应用深度卷积神经网络（DCNN）训练人体组织分割模型。使用训练好的模型对组织进行分割。根据相同人体组织的CT数值大致相同这一常识，通过分割可以生成无阴影伪影的模板图像，然后用特定组织的相应CT数值填充每个组织。从未校正图像中减去模板图像，生成带有图像细节和阴影伪影的残差图像。阴影伪影主要是低频信号，而图像细节主要是高频信号。因此，我们提出了一种自适应滤波器来准确分离阴影伪影和图像细节。最后，从原始图像中删除估计的阴影伪影以生成校正后的图像。
在Catphan©504研究中，校正后图像感兴趣区域（ROI）的CT数值误差从109降低到11 HU，图像对比度平均提高了1.46倍。在患者骨盆研究中，所选ROI的CT数值误差从198降低到10 HU。校正后，从ROI计算出的SNU从24%降至9%。
所提出的使用DCNN和AF的阴影校正方法可能在未来临床实践中找到有用的应用。
