Motion parameter decoupling and motion constraint-driven optimization for correcting rigid motion artifacts in cone-beam computed tomography

doi:10.12122/j.issn.1673-4254.2026.08.24

Abstract

Abstract:

Objective To solve the problem of rigid motion artifacts caused by patient movement and gantry vibration during long-duration cone-beam computed tomography (CBCT) using geometric parameter decoupling and motion-constrained optimization. Methods Using a motion estimation framework based on 3D-2D rigid registration, the rigid motion parameters were categorized into out-of-plane and in-plane motions, and a stepwise optimization sequence was designed to decouple the mutual interference among the parameters. In response to dynamic evolution of artifact characteristics from multi-contour overlap to edge blurring during the iterative process, a progressive cost function was formulated to facilitate an adaptive transition of the optimization objective from projection data consistency constraints to structural detail recovery. To address the issue of multi-solution in 3D-2D registration and the potential spatial misalignment introduced during iteration, a motion estimation constraint mechanism was incorporated to eliminate global bias, thereby enhancing the algorithm's convergence. Results The proposed algorithm accurately estimated rigid motion trajectories and restored the images via motion compensation. On head simulation data, this model achieved optimal quantitative metrics across 3 motion levels, improved PSNR by 2.1% and SSIM by 2.9%, and reduced RMSE by 6.5% compared to the suboptimal methods. Additional validation on knee simulation and real porcine data further demonstrated its efficacy and generalization capability. Conclusion The proposed rigid motion artifact correction algorithm demonstrates good performance in estimating motion trajectories and compensating for image artifacts, thus providing a viable and robust solution for suppressing rigid motion artifacts in clinical CBCT imaging.

Key words: rigid motion artifact correction, cone-beam computed tomography, 3D-2D rigid registration, parameter decoupling, motion constraint

Haotao JIANG, Yongbo WANG, Zhaoying BIAN. Motion parameter decoupling and motion constraint-driven optimization for correcting rigid motion artifacts in cone-beam computed tomography[J]. Journal of Southern Medical University, 2026, 46(8): 1967-1980.

Figures/Tables 16

Fig.1 Flow chart of the motion estimation method.

Fig.2 Schematic diagram of the coordinate system used in the experiment.

Fig.3 Flow chart of the proposed CBCT rigid motion artifact correction method.

Fig.4 Experimental setup. A: Xi'an Jiaotong University CBCT experimental platform. B: Porcine limb sample with controlled gimbal on CBCT rotating table.

Fig.5 Transverse-view results of different artifact-correction methods on head simulation data.

Fig.6 Coronal-view results of different artifact-correction methods on head simulation data.

Fig.7 Sagittal-view results of different artifact-correction methods on head simulation data.

Fig.8 Rigid motion estimation results of different motion estimation methods on head simulation data.

Tab.1 Quantitative indicators of different motion artifact correction algorithms on head simulation data

Method	Minor artifact			Moderate artifact			Severe artifact
Method	PSNR	RMSE	SSIM	PSNR	RMSE	SSIM	PSNR	RMSE	SSIM
Chen et al.	34.25	4.94	0.8709	31.20	7.02	0.8235	28.84	9.05	0.7800
Sun et al.	33.45	5.32	0.8517	29.98	7.53	0.7873	28.85	9.04	0.7575
LIT-Former	34.57	4.70	0.8613	29.54	7.98	0.7660	26.99	11.25	0.7037
Restormer	34.79	4.64	0.8670	30.69	7.20	0.8014	28.01	10.02	0.7359
TT-Unet	34.11	5.12	0.8565	30.09	7.41	0.7716	27.48	10.58	0.7134
Proposed	35.40	4.33	0.8934	31.49	6.78	0.8264	30.57	7.55	0.8018

Fig.9 Results of motion-estimation methods on knee simulation data.

Tab.2 Quantitative metrics of different motion estimation methods on knee simulation data

Methods	PSNR	RMSE	SSIM
Chen et al.	31.8366	6.5270	0.8881
Sun et al.	33.9899	5.1114	0.9174
Proposed	35.7197	4.1740	0.9531

Fig.10 Comparative results of different artifact correction methods on real data (axial and sagittal views).

Tab.3 Quantitative metrics of different motion estimation methods on real porcine data

Methods	PSNR	RMSE	SSIM
Chen et al.	33.0498	5.5761	0.8662
Sun et al.	31.7047	6.6268	0.8496
Proposed	35.5427	4.2599	0.8949

Fig.11 Comparative results of image restoration for motion estimation constraint validation.

Fig.12 Comparative results of rigid motion trajectory estimation for motion estimation constraint validation.

Fig.13 Comparative results of image restoration for staged cost function validation.

References 34

[1]	Spin-Neto R, Wenzel A. Patient movement and motion artefacts in cone beam computed tomography of the dentomaxillofacial region: a systematic literature review[J]. Oral Surg Oral Med Oral Pathol Oral Radiol, 2016, 121(4): 425-33. doi：10.1016/j.oooo.2015.11.019
[2]	Loubele M, Bogaerts R, Van Dijck E, et al. Comparison between effective radiation dose of CBCT and MSCT scanners for dento-maxillofacial applications[J]. Eur J Radiol, 2009, 71(3): 461-8. doi：10.1016/j.ejrad.2008.06.002
[3]	Nemtoi A, Czink C, Haba D, et al. Cone beam CT: a current overview of devices[J]. Dentomaxillofac Radiol, 2013, 42(8): 20120443. doi：10.1259/dmfr.20120443
[4]	Spin-Neto R, Costa C, Salgado DM, et al. Patient movement characteristics and the impact on CBCT image quality and interpretability[J]. Dentomaxillofac Radiol, 2018, 47(1): 20170216. doi：10.1259/dmfr.20170216
[5]	Nardi C, Taliani GG, Castellani A, et al. Repetition of examination due to motion artifacts in horizontal cone beam CT: comparison among three different kinds of head support[J]. J Int Soc Prev Community Dent, 2017, 7(4): 208-13. doi：10.4103/jispcd.jispcd_17_17
[6]	Hanzelka T, Dusek J, Ocasek F, et al. Movement of the patient and the cone beam computed tomography scanner: objectives and possible solutions[J]. Oral Surg Oral Med Oral Pathol Oral Radiol, 2013, 116(6): 769-73. doi：10.1016/j.oooo.2013.08.010
[7]	Malviya S, Voepel-Lewis T, Eldevik OP, et al. Sedation and general anaesthesia in children undergoing MRI and CT: adverse events and outcomes[J]. Br J Anaesth, 2000, 84(6): 743-8. doi：10.1093/oxfordjournals.bja.a013586
[8]	Beeres M, Wichmann JL, Paul J, et al. CT chest and gantry rotation time: does the rotation time influence image quality?[J]. Acta Radiol, 2015, 56(8): 950-4. doi：10.1177/0284185114544242
[9]	Berland LL, Smith JK. Multidetector-array CT: once again, technology creates new opportunities[J]. Radiology, 1998, 209(2): 327-9. doi：10.1148/radiology.209.2.9807555
[10]	Hernandez D, Eldib ME, Hegazy MAA, et al. A head motion estimation algorithm for motion artifact correction in dental CT imaging[J]. Phys Med Biol, 2018, 63(6): 065014. doi：10.1088/1361-6560/aab17e
[11]	Berger M, Maier A, Xia Y, et al. Motion compensated fan-beam CT by enforcing fourier properties of the sinogram[C]//2014 Third International Conference on Image Formation in X-ray Computed Tomography. Salt Lake City, Utah, USA. 2014: 329-32.
[12]	Zhang Y, Zhang LY. A rigid motion artifact reduction method for CT based on blind deconvolution[J]. Algorithms, 2019, 12(8): 155. doi：10.3390/a12080155
[13]	Hu ZL, Jiang CH, Zhang QY, et al. Wasserstein generative adversarial networks for motion artifact removal in dental CT imaging[C]//Medical Imaging 2019: Physics of Medical Imaging. February 16-21, 2019. San Diego, USA. SPIE, 2019: 113. doi：10.1117/12.2511818
[14]	Ali ASRA, Fusiello A, Landi C, et al. Motion artifacts detection in short-scan dental CBCT reconstructions[EB/OL]. 2023: arXiv: 2304.10154. . doi：10.1109/nssmicrtsd49126.2023.10337961
[15]	Weisenberger AG, Gleason SS, Goddard J, et al. A restraint-free small animal SPECT imaging system with motion tracking[J]. IEEE Trans Nucl Sci, 2005, 52(3): 638-44. doi：10.1109/tns.2005.851399
[16]	Herzog H, Tellmann L, Fulton R, et al. Motion artifact reduction on parametric PET images of neuroreceptor binding[J]. J Nucl Med, 2005, 46(6): 1059-65.
[17]	Kim JH, Nuyts J, Kyme A, et al. A rigid motion correction method for helical computed tomography (CT)[J]. Phys Med Biol, 2015, 60(5): 2047-73. doi：10.1088/0031-9155/60/5/2047
[18]	Kyme AZ, Se S, Meikle SR, et al. Markerless motion estimation for motion-compensated clinical brain imaging[J]. Phys Med Biol, 2018, 63(10): 105018. doi：10.1088/1361-6560/aabd48
[19]	Reaungamornrat S, Otake Y, Uneri A, et al. Tracker-on-C for cone-beam CT-guided surgery: evaluation of geometric accuracy and clinical applications[J]. Med Imag 2012 Image Guid Proced Rob Interv Model, 2012, 8316: 831609. doi：10.1117/12.911454
[20]	Ouadah S, Jacobson M, Stayman JW, et al. Correction of patient motion in cone-beam CT using 3D-2D registration[J]. Phys Med Biol, 2017, 62(23): 8813-31. doi：10.1088/1361-6560/aa9254
[21]	Yu HY, Wang G. Data consistency based rigid motion artifact reduction in fan-beam CT[J]. IEEE Trans Med Imaging, 2007, 26(2): 249-60. doi：10.1109/tmi.2006.889717
[22]	Berger M, Xia Y, Aichinger W, et al. Motion compensation for cone-beam CT using Fourier consistency conditions[J]. Phys Med Biol, 2017, 62(17): 7181-215. doi：10.1088/1361-6560/aa8129
[23]	Aichert A, Berger M, Wang J, et al. Epipolar consistency in transmission imaging[J]. IEEE Trans Med Imaging, 2015, 34(11): 2205-19. doi：10.1109/tmi.2015.2426417
[24]	Preuhs A, Maier A, Manhart M, et al. Double your views–exploiting symmetry in transmission imaging[M]//Medical Image Computing and Computer Assisted Intervention – MICCAI 2018. Cham: Springer International Publishing, 2018: 356-64. doi：10.1007/978-3-030-00928-1_41
[25]	Sun T, Kim JH, Fulton R, et al. An iterative projection-based motion estimation and compensation scheme for head X-ray CT[J]. Med Phys, 2016, 43(10): 5705. doi：10.1118/1.4963218
[26]	Berger M, Müller K, Aichert A, et al. Marker-free motion correction in weight-bearing cone-beam CT of the knee joint[J]. Med Phys, 2016, 43(3): 1235-48. doi：10.1118/1.4941012
[27]	Chen MY, He P, Feng P, et al. General rigid motion correction for computed tomography imaging based on locally linear embedding[J]. Opt Eng, 2018, 57(2): 1. doi：10.1117/1.oe.57.2.023102
[28]	Huang H, Siewerdsen JH, Zbijewski W, et al. Reference-free learning-based similarity metric for motion compensation in cone-beam CT[J]. Phys Med Biol, 2022, 67(12): 125020. doi：10.1088/1361-6560/ac749a
[29]	Sisniega A, Stayman JW, Yorkston J, et al. Motion compensation in extremity cone-beam CT using a penalized image sharpness criterion[J]. Phys Med Biol, 2017, 62(9): 3712-34. doi：10.1088/1361-6560/aa6869
[30]	Kingston A, Sakellariou A, Varslot T, et al. Reliable automatic alignment of tomographic projection data by passive auto-focus[J]. Med Phys, 2011, 38(9): 4934-45. doi：10.1118/1.3609096
[31]	Maur S, Stsepankou D, Hesser J. Auto-calibration by locally consistent contours for dental CBCT[J]. Phys Med Biol, 2018, 63(21): 215018. doi：10.1088/1361-6560/aae66d
[32]	Zamir SW, Arora A, Khan S, et al. Restormer: efficient transformer for high-resolution image restoration[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 18-24, 2022, New Orleans, LA, USA. IEEE, 2022: 5718-29. doi：10.1109/cvpr52688.2022.00564
[33]	Chen ZH, Niu C, Gao Q, et al. LIT-former: linking in-plane and through-plane transformers for simultaneous CT image denoising and deblurring[J]. IEEE Trans Med Imaging, 2024, 43(5): 1880-94. doi：10.1109/tmi.2024.3351723
[34]	Deng ZH, Zhang WK, Chen KL, et al. TT U-Net: temporal transformer U-Net for motion artifact reduction using PAD (pseudo all-phase clinical-dataset) in cardiac CT[J]. IEEE Trans Med Imaging, 2023, 42(12): 3805-16. doi：10.1109/tmi.2023.3310933