Xing Zhao

Dr. Xing Zhao


Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 June 2016 in PLOS ONE
Verified
Journal Article
Open access
PDF file
TLTing Luo
Ting Luo
CSChangrong Shi
Changrong Shi
XZXing Zhao
Xing Zhao
YZYunsong Zhao
Yunsong Zhao
JXJinqiu Xu
Jinqiu Xu
In this paper, we propose an automatic method of synthesizing panoramic radiographs from dental cone beam computed tomography (CBCT) data for directly observing the whole dentition without the superimposition of other structures. This method consists of three major steps. First, the dental arch curve is generated from the maximum intensity projection (MIP) of 3D CBCT data. Then, based on this curve, the long axial curves of the upper and lower teeth are extracted to create a 3D panoramic curved surface describing the whole dentition. Finally, the panoramic radiograph is synthesized by developing this 3D surface. Both open-bite shaped and closed-bite shaped dental CBCT datasets were applied in this study, and the resulting images were analyzed to evaluate the effectiveness of this method. With the proposed method, a single-slice panoramic radiograph can clearly and completely show the whole dentition without the blur and superimposition of other dental structures. Moreover, thickened panoramic radiographs can also be synthesized with increased slice thickness to show more features, such as the mandibular nerve canal. One feature of the proposed method is that it is automatically performed without human intervention. Another feature of the proposed method is that it requires thinner panoramic radiographs to show the whole dentition than those produced by other existing methods, which contributes to the clarity of the anatomical structures, including the enamel, dentine and pulp. In addition, this method can rapidly process common dental CBCT data. The speed and image quality of this method make it an attractive option for observing the whole dentition in a clinical setting.

15 June 2021 in Inverse Problems
Verified
Journal Article
Open access
SZShusen Zhao
Shusen Zhao
DXDimeng Xia
Dimeng Xia
XZXing Zhao
Xing Zhao
In planar objects computed tomography (CT), restricted to the scanning environment, projections can only be collected from limited angles. Moreover, limited by the emitting power of the x-ray source, only a few photons penetrate the long side of the planar objects, which results in the noise increasing in projections. Planar objects CT reconstruction based on these two conditions is mathematically corresponding to solving an ill-posed inverse problem. Although several iterative reconstruction algorithms of limited-angle CT were proposed, high-quality planar objects CT reconstruction algorithms with fast convergence are still the goals of many researchers. In order to address the aforementioned problems, we proposed a new optimization model for planar objects CT reconstruction. Inspired by the theory of ‘visible boundary and invisible boundary’ in limited-angle CT and the differentiation property of Fourier transform, a new optimization objective function is proposed in this paper. Based on the statistical noise model of existing CT system, the convex set constraint of the optimization model is given. Besides, the optimization model is solved by convex set projection and Fourier transform differentiation property. The proposed algorithm was evaluated with both simulated data and real data. The experimental results show that the proposed algorithm can achieve the effect of noise suppression, limited-angle artifacts reduction, and fast structure reconstruction when it applies to planar objects CT.

1 September 2021 in Medical Physics
Verified
Journal Article
Closed access
PDF file
WZWeibin Zhang
Weibin Zhang
SZShusen Zhao
Shusen Zhao
HPHuiying Pan
Huiying Pan
YZYunsong Zhao
Yunsong Zhao
XZXing Zhao
Xing Zhao

2 March 2021 in Physics in Medicine & Biology
Verified
Journal Article
Open access
PDF file
SZShusen Zhao
Shusen Zhao
HPHuiying Pan
Huiying Pan
WZWeibin Zhang
Weibin Zhang
DXDimeng Xia
Dimeng Xia
XZXing Zhao
Xing Zhao
In x-ray multispectral (or photon-counting) computed tomography (MCT), the object of interest is scanned under multiple x-ray spectra, and it can acquire more information about the scanned object than conventional CT, in which only one x-ray spectrum is used. The obtained polychromatic projections are utilized to perform material-selective and energy-selective image reconstruction. Compared with the conventional single spectral CT, MCT has a superior material distinguishability. Therefore, it has wide potential applications in both medical and industrial areas. However, the nonlinearity and ill condition of the MCT problem make it difficult to get high-quality and fast convergence of images for existing MCT reconstruction algorithms. In this paper, we proposed an iterative reconstruction algorithm based on an oblique projection modification technique (OPMT) for fast basis material decomposition of MCT. In the case of geometric inconsistency, along the current x-ray path, the oblique projection modification direction not only relates to the polychromatic projection equation of the known spectrum, but it also comprehensively refers to the polychromatic projection equation information of the unknown spectra. Moreover, the ray-by-ray correction makes it applicable to geometrically consistent projection data. One feature of the proposed algorithm is its fast convergence speed. The OPMT considers the information from multiple polychromatic projection equations, which greatly speeds up the convergence of MCT reconstructed images. Another feature of the proposed algorithm is its high flexibility. The ray-by-ray correction will be suitable for any common MCT scanning mode. The proposed algorithm is validated with numerical experiments from both simulated and real data. Compared with the ASD-NC-POCS and E-ART algorithms, the proposed algorithm achieved high-quality reconstructed images while accelerating the convergence speed of them.

20 November 2020 in IEEE Access
Verified
Journal Article
Open access
GMGenwei Ma
Genwei Ma
YZYining Zhu
Yining Zhu
XZXing Zhao
Xing Zhao
The x-ray computed tomography (CT) is essential for medical diagnosis and industrial nondestructive testing. The aim of CT is to recover or reconstruct image from projection data. However, in particular, the reconstructed image usually suffers from complex artifacts and noise, such as the sampling is insufficient or low-dose CT. In order to deal with such issues and achieve reconstruction, a full automatic reconstruction (FAR) net is proposed for CT reconstruction via deep learning technique. Different with the usual network in deep learning reconstruction, the proposed neural network is an end-to-end network by which the image is predicted directly from projection data. The main challenge for such a FAR-net is the space complexity of the CT reconstruction in full-connected (FC) network. For a CT image with the size N×N, a typical requirement of memory space for the image reconstruction is O(N4), for which is unacceptable by conventional calculation device, e.g. GPU workstation. In this paper, we utilize a series of smaller fully connected layers (FCL) to replace the huge Radon transform matrix based on the sparse nonnegative matrix factorization (SNMF) theory. By applying such an approach, the FAR-net is able to reconstruct images with the size 512×512 on only single workstation. The results of numerical experiments show that the projection matrix and the FAR-net is able to reconstruct the CT image from projection data with a superior quality to conventional methods such as optimization based approach. Meanwhile, the factorization for the inverse projection matrix is validated in simulation and real experiments.

5 September 2021 in Medical Physics
Verified
Journal Article
Closed access
PDF file
GMGenwei Ma
Genwei Ma
YZYinghui Zhang
Yinghui Zhang
XZXing Zhao
Xing Zhao
TWTong Wang
Tong Wang
HLHongwei Li
Hongwei Li