Abstract: Under the condition that the number of projection angles is constant, reducing the radiation dose under each angle is an effective way to realize low-dose CT. However, the reconstructed images obtained through this method can be very noisy. At present, the deep learning image denoising method represented by convolutional neural networks (CNN) has become a classical method for low-dose CT image denoising. Inspired by the good performance of transformer in computer vision tasks, this paper proposes a CNN transformer coupling network (CTC) to further improve the performance of CT image denoising. CTC network makes comprehensive use of local information association ability of CNN and global information capture ability of transformer, constructs eight core network blocks composed of CNN components and an improved transformer component, which are interconnected based on residual connection mechanism and information reuse mechanism. Compared with the existing four denoising networks, CTC network demonstrate better denoising ability and can realize high-precision low-dose CT image reconstruction.
QIAO Y Y, QIAO Z W. Low-dose CT image reconstruction method based on CNN and transformer coupling network[J]. CT Theory and Applications, 2022, 31(6): 697-707. DOI:10.15953/j.ctta.2022.114. (in Chinese)
Abstract: Reducing the dose of computed tomography (CT) is essential for reducing the radiation risk in clinical applications. With the rapid development and wide application of deep learning, it has brought new directions for the development of low-dose CT imaging algorithms. Unlike most existing prior-driven algorithms that benefit from manually designed prior functions or supervised learning schemes, in this paper, we use an energy-based deep model to learn the prior knowledge of normal-dose CT, and then in the iterative reconstruction phase, we integrate data consistency as a conditional item into the iterative generation model of low-dose CT, and realize the low-dose CT reconstruction through the prior experience of iterative updating training of Langevin dynamics. The experimental results show that the proposed method hold excellent noise reduction and detail retention capabilities.
ZHU Y Z, LV Q W, GUAN Y, et al. Low-dose CT reconstruction based on deep energy models[J]. CT Theory and Applications, 2022, 31(6): 709-720. DOI:10.15953/j.ctta.2021.077. (in Chinese)
Abstract: Low-dose computed tomography (CT) is a relatively safe method for disease screening. But low-dose CT images often contain severe noise and artifacts, which seriously affect the subsequent diagnosis. To solve this problem, this paper proposes a subspace projection and edge enhancement network (SPEENet). SPEENet hold an architecture of autoencoder, including two main modules: dual stream encoder and decoder. The dual stream encoder can be divided into two parts: noise image coding stream and edge information coding stream. The noise image coding stream removes the noise and artifacts in low-dose CT images by using the image features extracted from the low-dose CT images. The edge information coding stream mainly focuses on the edge information of low-dose CT images and fully utilize the edge information to preserve the structures. In order to make full use of the encoder features, this paper introduces the noise basis projection module to establish a basis based on the features of encoder and decoder, and uses this basis to project the features extracted by the encoder into the corresponding subspace to obtain better feature representation. In this paper, experiments are conducted on the public database to verify the effectiveness. The experimental results show that SPEENet can achieve better denoising performance than other low-dose CT denoising networks.
WEI Y L, YANG Z Y, XIA W J, et al. Low-dose CT denoising based on subspace projection and edge enhancement[J]. CT Theory and Applications, 2022, 31(6): 721-729. DOI:10.15953/j.ctta.2022.108. (in Chinese)
Abstract: Total variation (TV) minimum algorithm is an image reconstruction algorithm based on compressed sensing theory, which can realize the reconstruction of images with high accuracy from sparse projection or noisy projection data and has been widely used in computed tomography (CT), magnetic resonance imaging (MRI) and electronic paramagnetic resonance imaging (EPRI). Energy spectrum CT, T1 or T2 weighted MRI and EPRI both belong to multi-channel imaging. The channel-by-channel TV algorithm can achieve high-precision image reconstruction, but it ignores the similarity among the images of each channel while Nuclear TV algorithm is a TV algorithm that considers the image similarity among channels, and can realize high-precision image reconstruction. For multi-channel image reconstruction, taking CT reconstruction as an example, this paper proposes a nuclear TV multi-channel image reconstruction algorithm based on the framework of Chambolle-pock algorithm. Through the reconstruction experiments of simulated phantom and real CT image phantom, the accuracy of the algorithm is verified, the convergence of the algorithm is analyzed, the influence of algorithm parameters on the convergence rate is explored, and the sparse reconstruction ability and noisy projection reconstruction ability of the algorithm are evaluated. The experimental results show that the proposed algorithm can achieve higher reconstruction accuracy than the channel-by-channel TV algorithm. Nuclear TV algorithm is a high-precision multi-channel image reconstruction algorithm, which can be applied to multi-channel reconstruction of various imaging modes.
MA J Y, QIAO Z W. Nuclear TV multi-channel image reconstruction algorithms based on Chambolle-pock framework[J]. CT Theory and Applications, 2022, 31(6): 731-747. DOI:10.15953/j.ctta.2022.111. (in Chinese)
Abstract: Spectral computed tomography (CT) based on photon counting detector can simultaneously collect projection data of multiple spectral channels and obtain absorption characteristics of material within corresponding energy ranges, so it can be effectively applied to material identification and material decomposition. Principal component analysis is an excellent multivariate analysis technique, which can be applied to process multi-energy spectral CT data. In this paper, principal component analysis was performed on spectral CT data in projection domain and image domain respectively, and the analysis results were compared systematically. Meanwhile, in order to reduce the influence of noise and improve the color characterization performance of spectral CT images, the method of combining double domain filtering with pixel value square was proposed to denoise the noisy principal component images, and then the selected principal component images were mapped to RGB color channels. The experimental results demonstrate that the principal component analysis can obtain clear CT images and identify the different components of the substance, whether in the projection domain or the image domain. However, compared with the principal component analysis method in the image domain, principal component analysis in the projection domain can retain more details of the substance and acquire clearer color CT images.
DI Y X, KONG H H, NIU X W. Research on image analysis method of spectral CT based on principal component analysis[J]. CT Theory and Applications, 2022, 31(6): 749-760. DOI:10.15953/j.ctta.2022.077. (in Chinese)
Abstract: The cone-beam CT (CBCT) system based on the two-dimensional flat-panel detector technology is widely applied in patient location verification before radiotherapy. However, during the application of intraperitoneal tumor radiotherapy, severe shading and streaking artifacts caused by respiratory movement and intestinal peristalsis make it difficult to distinguish tumor areas from the CBCT images. Due to the non-rigid deformation of flexible organs such as the pancreas under the action of respiratory motion, it is hard to quantify deviation between the body surface motion monitoring results and the actual organ motion, and it is also difficult to monitor irregular motion represented by intestinal peristalsis. There is no effective solution to motion artifact correction in CBCT. Based on theory of biodynamics and common knowledge of human physiology, in this paper we propose a brand new radiotherapy image-guided cone-beam CT motion artifact correction method without motion monitoring or implantation of in-vivo markers. The proposed artifact correction strategy is designed based on the features of the artifact images and fusion of various CT image domain processing algorithms. The results suggest that the image quality of cone beam CT has been significantly improved after the application of this strategy in the clinical abdominal CBCT image processing. The average CT number error in typical soft tissue areas reduces from 90 HU to 30 HU, and the boundary of the intestinal cavity and surrounding soft tissue information are partially recovered. The proposed artifact correction strategy does not require respiratory gating or increase of projections, which can be integrated into existing workflows without marker implantation surgery. The motion-artifact-corrected CBCT images provide more accurate tumor localization information for image-guided radiotherapy of pancreatic carcinoma. The proposed method is proved practical and efficient for clinical applications
LUO C, REN Q, LIU J Q, et al. Development of motion artifact correction solutions for the cone-beam CT images during pancreatic cancer image-guided radiotherapy[J]. CT Theory and Applications, 2022, 31(6): 761-771. DOI:10.15953/j.ctta.2022.066. (in Chinese)
Abstract: In computed tomography (CT) imaging, artifacts will degrade the quality of reconstructed images. To solve this issue, in this paper we propose a method to remove concentric elliptical artifacts from CT images. This method is based on the idea of Directional total variation (DTV), which models the problem of elliptical artifact removal as an energy minimization problem, and establishes a variation-based model which is adapted the elliptical artifacts by analyzing the edge features of elliptical artifacts. Since the proposed model is a non-smooth convex optimization problem with a divisible structure, the alternating direction multiplier method (ADMM) is applied. Finally, the effectiveness of the model in removing elliptical artifacts is verified by simulation experiments.
LI Q X, JIN K, PANG Z F. A variational model for removing concentric elliptical artifacts from CT images[J]. CT Theory and Applications, 2022, 31(6): 773-781. DOI:10.15953/j.ctta.2022.085. (in Chinese)
Abstract: Rock and mineral samples contain a large amount of high-density metal substances, which often lead to metal artifacts in CT images and seriously affect the parameters analysis accuracy of rock and mineral samples. In order to improve the quality of CT images and suppress metal artifacts, in this paper we propose a dual-domain adaptive network-based metal artifact reduction algorithm for CT images of rock samples (DDA-CNN-MAR). The metal artifacts are suppressed by the projection domain network and the image domain network successively, and the dual domain processing results are adaptively fused to realize the end-to-end mapping from images with artifacts to artifact-free images. The algorithm is based on the residual encoder-decoder network model (RED-CNN), which is easy to extract features and restore image details. The dual-domain structure can adaptively adjust the weights of the projection domain (artifact suppression) and image domain (detail inpainting) to obtain optimal reduction results. The experimental results show that, compared with the RED-CNN metal artifact denoising method in the image domain, the MSE of the image corrected by the DDA-CNN-MAR method is reduced by 2.570, while the PSNR and SSIM are increased by 1.218 dB and 0.018 respectively, which effectively improves the CT image quality.
ZUO S J, FENG P, HUANG P, et al. Metal artifact reduction algorithm for CT images of rock and mineral samples based on dual-domain adaptive network[J]. CT Theory and Applications, 2022, 31(6): 783-792. DOI:10.15953/j.ctta.2022.041. (in Chinese)
Abstract: The traditional ultrasonic dynamic focusing imaging which simply uses the delay-and-sum method in beamforming shows low resolution and poor contrast. In this study, based on the coherent pixel-based beamforming (Coherent PB), a novel Wiener pre-filter with high computational efficiency is applied to correct the phase of the entire array signal by using the pulse-echo at the focal point. We also apply a Wiener post-filter calculating adaptive weights for the beamforming results of each sub-aperture to suppress the noise and artifacts. The proposed method's effectiveness is verified through simulation experiments, phantom experiments, and in vivo experiments. Compared with Coherent PB, the proposed method significantly improves the axial resolution and contrast of images while maintaining the high lateral resolution and computational efficiency, which shows certain clinical application value.
PU S L, XIE H W, GUO H, et al. Coherent beam-forming combined with wiener filter in ultrasound imaging[J]. CT Theory and Applications, 2022, 31(6): 793-808. DOI:10.15953/j.ctta.2022.043. (in Chinese)
Abstract: Blood flow is an important physiological parameter of the human body. Real-time measurement of blood flow in the brain, skeletal muscle, and breast tissue is of great significance for disease diagnosis, treatment, surgery, and intensive care. Near-Infrared Diffuse Correlation Spectroscopy (DCS) is a new-type tissue blood flow measurement technology. When using DCS technology for blood flow measurement, the light source-detector (S-D) at each distance contains different degrees of mixed signals of superficial and deep tissues, among which the superficial signals show greater impact on the extraction of blood flow in deep tissues. This paper combines the Nth order linear algorithm (NL algorithm) with the independent component analysis algorithm (Independent Component Analysis, ICA) to separate and process the short-range and long-range optical signals obtained by DCS technology. The computer simulation shows that the algorithm proposed in this paper can better separate the blood flow signals of the superficial and deep tissues, and demonstrates important potential for the application of DCS technology in clinical blood flow measurement in the future.
SUN L, BAI J, DI W Q, et al. Research on blood flow separation algorithm of diffuse light correlation spectrum based on ICA[J]. CT Theory and Applications, 2022, 31(6): 809-820. DOI:10.15953/j.ctta.2022.132. (in Chinese)