State Key Laboratory of Dynamic Testing Technology, North University of China, Taiyuan, China.
School of Information and Communication Engineering, North University of China, Taiyuan, China.
J Xray Sci Technol. 2022;30(6):1229-1242. doi: 10.3233/XST-221233.
Low-dose computed tomography (LDCT) is an effective method for reducing radiation exposure. However, reducing radiation dose leads to considerable noise in the reconstructed image that can affect doctor's judgment.
To solve this problem, this study proposes a local total variation and improved wavelet residual convolutional neural network (LTV-WRCNN) denoising model.
The model first introduces local total variation (LTV) to decompose the LDCT image into cartoon and texture image. Next, the texture image is filtered using the non-local mean (NLM). Then, the cartoon image is added to the filtered texture image to obtain the preprocessing image. Finally, the pre-processed image is fed into the improved wavelet residual neural network (WRCNN) to obtain an improved image. Additionally, we also introduce a compound loss in wavelet domain that combines mean squared error loss and directional regularization loss to separate the structural details from noise more thoroughly.
Compared with state-of-the-art methods, the peak-signal-to-noise ratio (PSNR) value and the structure similarity (SSIM) value of the processed CT images using the new proposed model are 33.4229 dB and 0.9158. Study also shows that applying new model obtains better results visually and numerically, especially in terms of the preservation of structural details.
The proposed new model is feasible and effective in improving the quality of LDCT images.
低剂量计算机断层扫描（LDCT）是降低辐射暴露的有效方法。然而，降低辐射剂量会导致重建图像中产生相当大的噪声，从而影响医生的判断。
为了解决这个问题，本研究提出了一种局部全变差和改进的小波残差卷积神经网络（LTV-WRCNN）去噪模型。
该模型首先引入局部全变差（LTV）将 LDCT 图像分解为卡通和纹理图像。接下来，使用非局部均值（NLM）对纹理图像进行滤波。然后，将卡通图像添加到滤波后的纹理图像中，得到预处理图像。最后，将预处理后的图像输入到改进的小波残差神经网络（WRCNN）中，得到改进后的图像。此外，我们还在小波域中引入了一种复合损失，将均方误差损失和方向正则化损失结合起来，以更彻底地从噪声中分离结构细节。
与最先进的方法相比，使用新提出的模型处理后的 CT 图像的峰值信噪比（PSNR）值和结构相似性（SSIM）值分别为 33.4229 dB 和 0.9158。研究还表明，新模型在视觉和数值上都能获得更好的效果，特别是在结构细节的保留方面。
所提出的新模型在提高 LDCT 图像质量方面是可行和有效的。
