Improve angular resolution for sparse-view CT with residual convolutional neural network

Liang, Kaichao; Yang, Honggang; Kang, Kejun; Xing, Yuxiang

doi:10.1117/12.2293319

Cited by 26 publications

(17 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While those approaches exploit the deep-neural-network in the image domain of CT rather than in the sinogram domain, this paper focuses on restoring the missing data in the sinogram domain so that one can reconstruct images by use of the well-established reconstruction algorithms in practical uses. There is another approach using deep neural network for synthesizing sinogram [37]. This method trains a network to learn residual between input sinogram and sparsely sampled sinograms from different angular directions, and concatenates those multiple sparselysampled sinograms to generate a full-view sinogram with an optimization network.…”

mentioning

confidence: 99%

Deep-Neural-Network-Based Sinogram Synthesis for Sparse-View CT Image Reconstruction

Lee

Kim

et al. 2019

IEEE Trans. Radiat. Plasma Med. Sci.

207

154

View full text Add to dashboard Cite

Recently, a number of approaches to low-dose computed tomography (CT) have been developed and deployed in commercialized CT scanners. Tube current reduction is perhaps the most actively explored technology with advanced image reconstruction algorithms. Sparse data sampling is another viable option to the low-dose CT, and sparse-view CT has been particularly of interest among the researchers in CT community. Since analytic image reconstruction algorithms would lead to severe image artifacts, various iterative algorithms have been developed for reconstructing images from sparsely view-sampled projection data. However, iterative algorithms take much longer computation time than the analytic algorithms, and images are usually prone to different types of image artifacts that heavily depend on the reconstruction parameters. Interpolation methods have also been utilized to fill the missing data in the sinogram of sparse-view CT thus providing synthetically full data for analytic image reconstruction. In this paper, we introduce a deep-neuralnetwork-enabled sinogram synthesis method for sparse-view CT, and show its outperformance to the existing interpolation methods and also to the iterative image reconstruction approach.Index Terms-Deep learning, low-dose computed tomography (CT), sparse-view CT, view interpolation. I. INTRODUCTIONW ITH increased use of X-ray computed tomography (CT) in clinics, potential radiation hazard has been alarmed [1], [2]. There have been developed a host

show abstract

mentioning

confidence: 99%

Deep-Neural-Network-Based Sinogram Synthesis for Sparse-View CT Image Reconstruction

Lee

Kim

et al. 2019

IEEE Trans. Radiat. Plasma Med. Sci.

207

154

View full text Add to dashboard Cite

show abstract

“…The projection estimation network is to estimate missing projection data from sparse view projections p sp to gain complete data [23]. As Fig.…”

Section: Architectures Of the Networkmentioning

confidence: 99%

Comparison of projection domain, image domain, and comprehensive deep learning for sparse-view X-ray CT image reconstruction

Liang,

Yang,

Xing

2018

Preprint

Self Cite

View full text Add to dashboard Cite

X-ray Computed Tomography (CT) imaging has been widely used in clinical diagnosis, non-destructive examination, and public safety inspection. Sparse-view (sparse view) CT has great potential in radiation dose reduction and scan acceleration. However, sparse view CT data is insufficient and traditional reconstruction results in severe streaking artifacts. In this work, based on deep learning, we compared image reconstruction performance for sparse view CT reconstruction with projection domain network, image domain network, and comprehensive network combining projection and image domains. Our study is executed with numerical simulated projection of CT images from real scans. Results demonstrated deep learning networks can effectively reconstruct rich high frequency structural information without streaking artefact commonly seen in sparse view CT. A comprehensive network combining deep learning in both projection domain and image domain can get best results.

show abstract

“…The sinogram inpainting problem is similar to the image inpainting problem which is based on the information in the image to restore the missing parts of the image. Some studies showed that DNNs have the potential to extract sinogram information effectively, to repair missing projection data and improve the image quality in the sparse angle reconstruction [31,32,33].…”

Section: Introductionmentioning

confidence: 99%

Promising Generative Adversarial Network Based Sinogram Inpainting Method for Ultra-Limited-Angle Computed Tomography Imaging

Cai

Wang

et al. 2019

Sensors

View full text Add to dashboard Cite

Limited-angle computed tomography (CT) image reconstruction is a challenging problem in the field of CT imaging. In some special applications, limited by the geometric space and mechanical structure of the imaging system, projections can only be collected with a scanning range of less than 90°. We call this kind of serious limited-angle problem the ultra-limited-angle problem, which is difficult to effectively alleviate by traditional iterative reconstruction algorithms. With the development of deep learning, the generative adversarial network (GAN) performs well in image inpainting tasks and can add effective image information to restore missing parts of an image. In this study, given the characteristic of GAN to generate missing information, the sinogram-inpainting-GAN (SI-GAN) is proposed to restore missing sinogram data to suppress the singularity of the truncated sinogram for ultra-limited-angle reconstruction. We propose the U-Net generator and patch-design discriminator in SI-GAN to make the network suitable for standard medical CT images. Furthermore, we propose a joint projection domain and image domain loss function, in which the weighted image domain loss can be added by the back-projection operation. Then, by inputting a paired limited-angle/180° sinogram into the network for training, we can obtain the trained model, which has extracted the continuity feature of sinogram data. Finally, the classic CT reconstruction method is used to reconstruct the images after obtaining the estimated sinograms. The simulation studies and actual data experiments indicate that the proposed method performed well to reduce the serious artifacts caused by ultra-limited-angle scanning.

show abstract

Improve angular resolution for sparse-view CT with residual convolutional neural network

Cited by 26 publications

References 10 publications

Deep-Neural-Network-Based Sinogram Synthesis for Sparse-View CT Image Reconstruction

Deep-Neural-Network-Based Sinogram Synthesis for Sparse-View CT Image Reconstruction

Comparison of projection domain, image domain, and comprehensive deep learning for sparse-view X-ray CT image reconstruction

Promising Generative Adversarial Network Based Sinogram Inpainting Method for Ultra-Limited-Angle Computed Tomography Imaging

Contact Info

Product

Resources

About