Iterative model-based image reconstruction algorithms in cone beam computed tomography (CBCT) require repetitive forward and backward projection operations. We compare the quality of the branchless distance-driven (BDD) projector in iterative CBCT reconstruction with ray- and voxel-based methods in both regular and low-dose examinations, and introduce a hybrid approach that aims at faster computation by using the BDD as the backprojector only. We also demonstrate the potential of the BDD in FDK reconstructions.
Approach
Two measured and one simulated datasets are used. Contrast-to-noise ratio (CNR) and modulation transfer function values are computed for one measured dataset. The structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR) are computed with the simulated data.
Results
Based on our results, BDD has reduced noise and better CNR compared to the other methods, with the quality dependent on the scanner geometry. The CNR is improved by 21% with the BDD and 4% with the hybrid method. BDD improves SSIM values by approximately 3.3% in the lowest dose case and 1% in the highest dose case, while for PSNR the values are 5% - 10% better. For the hybrid method, the SSIM improvements range from 0.8% - 2.2%, and the PSNR from 3.7% - 6.6 %. The hybrid method with the BDD as a backprojector can be computationally twice faster with similar image quality.
Conclusions
The hybrid projector is a good choice as a compromise between image quality and computation time. Furthermore, BDD and the hybrid projector are better choices in low-dose CBCT reconstructions.
HankeRFuchsTUhlmannN. X-ray based methods for non-destructive testing and material characterization. Nuclear Inst Methods Phys Res Sec A: Acceler Spect Detect Assoc Equipment2008; 591: 14–18.
2.
EvansLMargettsLCasalegnoV, et al.Transient thermal finite element analysis of CFC–cu ITER monoblock using X-ray tomography data. Fusion Eng Design2015; 100: 100–111.
3.
DambrogioJGhassaeiASmithDS, et al.Unlocking history through automated virtual unfolding of sealed documents imaged by X-ray microtomography. Nat Commun2021; 12: 1184.
4.
SealesWBParkerCSSegalM, et al.From damage to discovery via virtual unwrapping: Reading the scroll from en-gedi. Sci Adv2016; 2: e1601247.
5.
TonaiSKuboYTsangMY, et al.A new method for quality control of geological cores by X-ray computed tomography: Application in IODP expedition 370. Front Earth Sci2019; 7. DOI: 10.3389/feart.2019.00117.
6.
CastañónDABabaheidarianP. Joint reconstruction and material classification in spectral CT. In: Ashok A, Neifeld MA, Gehm ME et al. (eds) Anomaly detection and imaging with X-Rays (ADIX) III. SPIE. DOI:10.1117/12.2309663.
7.
MegherbiNFlittonGTBreckonTP. A classifier based approach for the detection of potential threats in CT based baggage screening. In: 2010 IEEE International conference on image processing. IEEE. DOI:10.1109/icip.2010.5653676.
8.
WangG. X-ray micro-CT with a displaced detector array. Med Phys2002; 29: 1634–1636.
9.
De ManBBasuS. Distance-driven projection and backprojection. In: 2002 IEEE Nuclear science symposium conference record, Vol. 3, pp.1477–1480. IEEE. DOI:10.1109/NSSMIC.2002.1239600.
10.
JiaXLouYLewisJ, et al.GPU-based fast low-dose cone beam CT reconstruction via total variation. J X-Ray Sci Technol2011; 19: 139–154.
11.
ParkHGShinYGLeeH. A fully GPU-based ray-driven backprojector via a ray-culling scheme with voxel-level parallelization for cone-beam CT reconstruction. Technol Cancer Res Treat2015; 14: 709–720.
12.
LongYFesslerJABalterJM. 3D forward and back-projection for X-ray CT using separable footprints. IEEE Trans Med Imag2010; 29: 1839–1850.
13.
WettenhoviVVVauhkonenMKolehmainenV. OMEGA—open-source emission tomography software. Phys Med Biol2021; 66: 065010.
14.
ZengGGullbergG. A ray-driven backprojector for backprojection filtering and filtered backprojection algorithms. In: 1993 IEEE conference record nuclear science symposium and medical imaging conference. NSSMIC-93, 1993, IEEE. DOI:10.1109/nssmic.1993.701833.
15.
ZwickerMPfisterHvan BaarJ, et al.Ewa splatting. IEEE Trans Visualiz Comput Graphics2002; 8: 223–238.
16.
FesslerJA. Penalized weighted least-squares image reconstruction for positron emission tomography. IEEE Trans Med Imaging1994; 13: 290–300.
17.
LiuRHeLLuoY, et al.Singular value decomposition-based 2D image reconstruction for computed tomography. J X-Ray Sci Technol2017; 25: 113–134.
18.
LougovskiAHofheinzFMausJ, et al.A volume of intersection approach for on-the-fly system matrix calculation in 3D PET image reconstruction. Phys Med Biol2014; 59: 561–577.
YuHWangG. Finite detector based projection model for high spatial resolution. J X-Ray Sci Technol2012; 20: 229–238.
21.
ManBDBasuS. Distance-driven projection and backprojection in three dimensions. Phys Med Biol2004; 49: 2463–2475.
22.
LiuRFuLDe ManB, et al.GPU-based branchless distance-driven projection and backprojection. IEEE Trans Comput Imag2017; 3: 617–632.
23.
SchlifskeDMedeirosH. A fast GPU-based approach to branchless distance-driven projection and back-projection in cone beam CT. In: Kontos D, Flohr TG and Lo JY (eds) Medical imaging 2016: Physics of medical imaging. SPIE. DOI:10.1117/12.2216628.
24.
MitraAPolitteDGWhitingBR, et al.Multi-GPU acceleration of branchless distance driven projection and backprojection for clinical helical CT. J Imag Sci Technol2017; 61: 010405.
25.
XieXMcGaffinMGLongY, et al.Accelerating separable footprint (SF) forward and back projection on GPU. In: Flohr TG, Lo JY and Gilat Schmidt T (eds) SPIE Proceedings, 2017. SPIE. DOI:10.1117/12.2252010.
26.
BasuSDe ManB. Branchless distance driven projection and backprojection. In: Bouman CA, Miller EL and Pollak I (eds) SPIE Proceedings, 2006. SPIE. DOI:10.1117/12.659893.
ChambolleAPockT. A first-order primal-dual algorithm for convex problems with applications to imaging. J Math Imag Vision2011; 40: 120–145.
29.
SidkyEYJørgensenJHPanX. Convex optimization problem prototyping for image reconstruction in computed tomography with the Chambolle–Pock algorithm. Phys Med Biol2012; 57: 3065–3091.
30.
CondatL. A primal-dual splitting method for convex optimization involving lipschitzian, proximable and linear composite terms. J Optim Theory Appl2013; 158: 460–479.
31.
VũBC. A splitting algorithm for dual monotone inclusions involving cocoercive operators. Adv Comput Math2011; 38: 667–681.
32.
BuadesACollBMorelJM. A review of image denoising algorithms, with a new one. Mult Model Sim2005; 4: 490–530.
33.
WettenhoviVVHietanenANiinimäkiK, et al.Comparison of three different projectors for cone beam CT. In: 2022 IEEE Nuclear science symposium and medical imaging conference (NSS/MIC), 2022. IEEE. DOI:10.1109/nss/mic44845.2022.10399001.
34.
GaligekereRWiesentKHoldsworthD. Cone-beam reprojection using projection-matrices. IEEE Trans Med Imag2003; 22: 1202–1214.
35.
JiaXDongBLouY, et al.GPU-based iterative cone-beam CT reconstruction using tight frame regularization. Phys Med Biol2011; 56: 3787–3807.
36.
van AarleWPalenstijnWJBeenhouwerJD, et al.The ASTRA toolbox: A platform for advanced algorithm development in electron tomography. Ultramicroscopy2015; 157: 35–47.
37.
BiguriADosanjhMHancockS, et al.TIGRE: a MATLAB-GPU toolbox for CBCT image reconstruction. Biomed Phys Eng Express2016; 2: 055010.
38.
ZengGGullbergG. Unmatched projector/backprojector pairs in an iterative reconstruction algorithm. IEEE Trans Med Imag2000; 19: 548–555.
39.
KamphuisCBeekmanFJvan RijkPP, et al.Dual matrix ordered subsets reconstruction for accelerated 3d scatter compensation in single-photon emission tomography. European J Nucl Med Mol Imag1997; 25: 8–18.
40.
ZengGL. Counter examples for unmatched projector/backprojector in an iterative algorithm. Chinese J Acad Radiol2019; 1: 13–24.
41.
ReidEJDrummyLFBoumanCA, et al.Multi-resolution data fusion for super resolution imaging. IEEE Trans Comput Imag2022; 8: 81–95.
ChouzenouxEPesquetJCRiddellC, et al.Convergence of proximal gradient algorithm in the presence of adjoint mismatch *. Inverse Prob2021; 37: 065009.
44.
ChouzenouxEContrerasAPesquetJC, et al.Convergence results for primal-dual algorithms in the presence of adjoint mismatch. SIAM J Imag Sci2023; 16: 1–34.
45.
DongYHansenPCHochstenbachME, et al.Fixing nonconvergence of algebraic iterative reconstruction with an unmatched backprojector. SIAM J Sci Comput2019; 41: A1822–A1839.
HahnKSchöndubeHStierstorferK, et al.A comparison of linear interpolation models for iterative CT reconstruction. Med Phys2016; 43: 6455–6473.
48.
JosephPM. An improved algorithm for reprojecting rays through pixel images. IEEE Trans Med Imag1982; 1: 192–196.
49.
SiddonRL. Fast calculation of the exact radiological path for a three-dimensional CT array. Med Phys1985; 12: 252–255.
50.
JacobsFSundermannESutterBD, et al.A fast algorithm to calculate the exact radiological path through a pixel or voxel space. J Comput Inf Technol1998; 6: 89–94.
51.
ChristiaensMDe SutterBDe BosschereK, et al.A fast, cache-aware algorithm for the calculation of radiological paths exploiting subword parallelism. J Syst Archit1999; 45: 781–790.
52.
StoneJEGoharaDShiG. OpenCL: A parallel programming standard for heterogeneous computing systems. Comput Sci Eng2010; 12: 66–73.
53.
AlpayAHeuvelineV. SYCL beyond OpenCL: The architecture, current state and future direction of hipSYCL. In: Proceedings of the international workshop on openCL. IWOCL ’20, ACM. DOI:10.1145/3388333.3388658.
54.
LewisJ. Fast template matching. Vis Interface1994; 95: 120–123.
55.
HensleyJScheuermannTCoombeG, et al.Fast summed-area table generation and its applications. Comput Graph Forum2005; 24: 547–555.
56.
NehabDMaximoALimaRS, et al.GPU-efficient recursive filtering and summed-area tables. ACM Trans Graphics2011; 30: 1–12.
57.
YalamanchiliPArshadUMohammedZ, et al.ArrayFire - A high performance software library for parallel computing with an easy-to-use API, 2015. https://github.com/arrayfire/arrayfire.
58.
ChambolleAEhrhardtMJRichtárikP, et al.Stochastic primal-dual hybrid gradient algorithm with arbitrary sampling and imaging applications. SIAM J Optim2018; 28: 2783–2808.
59.
TangJEhrhardtMSchönliebCB. Stochastic primal-dual three operator splitting with arbitrary sampling and preconditioning. arXiv 2022. DOI: 10.48550/ARXIV.2208.01631. 2208.01631.
60.
BeckATeboulleM. A fast iterative shrinkage-thresholding algorithm for linear inverse problems. SIAM J Imag Sci2009; 2: 183–202.
61.
ZhangHZengDZhangH, et al.Applications of nonlocal means algorithm in low-dose X-ray CT image processing and reconstruction: A review. Med Phys2017; 44: 1168–1185.
62.
ChenYGaoDNieC, et al.Bayesian statistical reconstruction for low-dose x-ray computed tomography using an adaptive-weighting nonlocal prior. Comput Med Imag Graphics2009; 33: 495–500.
63.
ZhangHMaJWangJ, et al.Statistical image reconstruction for low-dose CT using nonlocal means-based regularization. Comput Med Imag Graphics2014; 38: 423–435.
64.
KelmZSBlezekDBartholmaiB, et al.Optimizing non-local means for denoising low dose CT. In: 2009 IEEE international symposium on biomedical imaging: from nano to macro, 2009. IEEE. DOI:10.1109/isbi.2009.5193134.
65.
GiraldoJCRKelmZSGuimaraesLS, et al.Comparative study of two image space noise reduction methods for computed tomography: Bilateral filter and nonlocal means. In: 2009 Annual international conference of the ieee engineering in medicine and biology society, 2009. pp.3529–3532. DOI:10.1109/IEMBS.2009.5334714.
66.
LiZYuLTrzaskoJD, et al.Adaptive nonlocal means filtering based on local noise level for CT denoising. Med Phys2014; 41: 011908.
67.
ChenYYangZHuY, et al.Thoracic low-dose CT image processing using an artifact suppressed large-scale nonlocal means. Phys Med Biol2012; 57: 2667.
68.
SegarsWPSturgeonGMendoncaS, et al.4D XCAT phantom for multimodality imaging research. Med Phys2010; 37: 4902–4915.
69.
BadalABadanoA. Accelerating monte carlo simulations of photon transport in a voxelized geometry using a massively parallel graphics processing unit. Med Phys2009; 36: 4878–4880.
70.
LiTFengHXuZ, et al.Comparison of different analytical edge spread function models for MTF calculation using curve-fitting. In: Maître H, Sun H, Lei B et al. (eds) MIPPR 2009: Remote Sensing and GIS data processing and other applications, 2009. SPIE. DOI:10.1117/12.832793.
71.
WangZBovikACSheikhHR, et al.Image quality assessment: From error measurement to structural similarity. IEEE Trans Image Process2004; 13: 600–612.
72.
WangZBovikAC. Mean squared error: Love it or leave it?IEEE Signal Process Mag2009; 26: 98–117.
