A low-dose CT reconstruction method using sub-pixel anisotropic diffusion

doi:10.12122/j.issn.1673-4254.2025.01.19

Abstract

Abstract:

Objective We present a new low-dose CT reconstruction method using sub-pixel and anisotropic diffusion. Methods The sub-pixel intensity values and their second-order differences were obtained using linear interpolation techniques, and the new gradient information was then embedded into an anisotropic diffusion process, which was introduced into a penalty-weighted least squares model to reduce the noise in low-dose CT projection data. The high-quality CT image was finally reconstructed using the classical filtered back-projection (FBP) algorithm from the estimated data. Results In the Shepp-Logan phantom experiments, the structural similarity (SSIM) index of the CT image reconstructed by the proposed algorithm, as compared with FBP, PWLS-Gibbs and PWLS-TV algorithms, was increased by 28.13%, 5.49%, and 0.91%, the feature similarity (FSIM) index was increased by 21.08%, 1.78%, and 1.36%, and the root mean square error (RMSE) was reduced by 69.59%, 18.96%, and 3.90%, respectively. In the digital XCAT phantom experiments, the SSIM index of the CT image reconstructed by the proposed algorithm, as compared with FBP, PWLS-Gibbs and PWLS-TV algorithms, was increased by 14.24%, 1.43% and 7.89%, the FSIM index was increased by 9.61%, 1.78% and 5.66%, and the RMSE was reduced by 26.88%, 9.41% and 18.39%, respectively. In clinical experiments, the SSIM index of the image reconstructed using the proposed algorithm was increased by 19.24%, 15.63% and 3.68%, the FSIM index was increased by 4.30%, 2.92% and 0.43%, and the RMSE was reduced by 44.60%, 36.84% and 15.22% in comparison with FBP, PWLS-Gibbs and PWLS-TV algorithms, respectively. Conclusion The proposed method can effectively reduce the noises and artifacts while maintaining the structural details in low-dose CT images.

Key words: low-dose computed tomography, anisotropic diffusion, sub-pixel, image reconstruction

Shizhou TANG, Ruolan SU, Shuting LI, Zhenzhen LAI, Jinhong HUANG, Shanzhou NIU. A low-dose CT reconstruction method using sub-pixel anisotropic diffusion[J]. Journal of Southern Medical University, 2025, 45(1): 162-169.

Figures/Tables 13

Fig.1 Sub-pixel under different distance hs: (A)hs1; (B)hs1.

Fig.2 Digital phantom: Shepp-Logan phantom (A) and XCAT phantom (B).

Fig.3 Shepp-Logan images reconstructed by FBP method (A), PWLS-Gibbs method (B), PWLS-TV method (C) and the proposed PWLS-SPAD method (D).

Tab.1 Evaluation indexes of the reconstructed Shepp-Logan images by different methods

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.752083	0.913535	0.954934	0.963646
FSIM	0.767575	0.913088	0.916857	0.929354
RMSE	0.111569	0.041859	0.035302	0.033924

Fig.4 Zoomed-in views of the reconstructed Shepp-Logan phantom images by FBP method (A), PWLS-Gibbs method (B), PWLS-TV method (C) and the proposed PWLS-SPAD method (D).

Tab.2 Evaluation indexes of the region-of-interest （ROI） in Fig.4

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.203666	0.813477	0.850652	0.888047
FSIM	0.399392	0.790398	0.827324	0.853895

Fig.5 XCAT phantom images reconstructed by FBP method (A), PWLS-Gibbs method (B), PWLS-TV method (C) and the proposed PWLS-SPAD method (D).

Tab.3 Evaluation indexes of reconstructed XCAT results by different methods

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.793550	0.893789	0.840232	0.906532
FSIM	0.814784	0.877465	0.845212	0.893061
RMSE	0.073716	0.059505	0.066048	0.053903

Fig.6 Zoomed-in views of the reconstructed XCAT phantom images by FBP method (A), PWLS-Gibbs method (B), PWLS-TV method (C) and the proposed PWLS-SPAD method (D).

Tab.4 Evaluation indexes of ROI in Fig.6

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.748943	0.852549	0.798578	0.873204
FSIM	0.806244	0.862104	0.832281	0.867679

Tab.5 Evaluation indexes of the reconstructed low-dose clinical CT images by different methods

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.795249	0.820050	0.914574	0.948223
FSIM	0.936031	0.948504	0.972051	0.976237
RMSE	0.068670	0.060228	0.044872	0.038042
PSNR	23.26	24.40	26.96	28.39

Fig.8 Zoomed-in views of the ROI in the reconstructed images by FBP method at 400 mAs (A) and 50 mAs (B), by PWLS-Gibbs at 50 mAs (C), by PWLS-TV at 50 mAs (D), and by the proposed PWLS-SPAD method at 50 mAs (E).

Tab.6 Evaluation indexes of ROI in Fig.8

Index	FBP	PWLS-Gibbs	PWLS-TV	PWLS-SPAD
SSIM	0.622500	0.786073	0.783700	0.810993
FSIM	0.840642	0.865528	0.882331	0.919254

References 32

1	Vliegenthart R, Fouras A, Jacobs C, et al. Innovations in thoracic imaging: CT, radiomics, AI and X-ray velocimetry[J]. Respirology, 2022, 27(10): 818-33.
2	Bonney A, Malouf R, Marchal C, et al. Impact of low-dose computed tomography (LDCT) screening on lung cancer-related mortality[J]. Cochrane Database Syst Rev, 2022, 8(8): CD013829.
3	Gillies RJ, Schabath MB. Radiomics improves cancer screening and early detection[J]. Cancer Epidemiol Biomark Prev, 2020, 29(12): 2556-67.
4	牛善洲, 唐诗洲, 黄舒彦, et al. 基于高维PDE投影恢复的低剂量CT重建方法 [J]. 南方医科大学学报, 2024, 44(4): 682-8.
5	Zhang H, Wang J, Zeng D, et al. Regularization strategies in statistical image reconstruction of low-dose X-ray CT: a review[J]. Med Phys, 2018, 45(10): e886-e907.
6	邸江磊, et al. 基于深度学习的稀疏或有限角度CT重建方法研究综述 [J]. 激光与光电子学进展, 2023, 60(8): 42-79.
7	牛善洲, 梁礼境, 李硕, et al. 基于低维流形先验的低剂量CT重建方法[J]. 计算机工程与应用, 2023, 59(18): 242-8.
8	李进丹, 陈龙, 杨聪慧, et al. 低剂量PET/CT及PET/MRI临床研究进展[J]. 中国医学影像技术, 2021, 37(11): 1740-3.
9	陈世宣, 曾栋, 边兆英, et al. 基于联邦特征学习的多机型低剂量CT重建算法 [J]. 南方医科大学学报, 2024, 44(2): 333-43.
10	Yang SH, Pu Q, Lei CT, et al. Low-dose CT denoising with a high-level feature refinement and dynamic convolution network[J]. Med Phys, 2023, 50(6): 3597-611.
11	Wu DF, Kim K, El Fakhri G, et al. Iterative low-dose CT reconstruction with priors trained by artificial neural network[J]. IEEE Trans Med Imaging, 2017, 36(12): 2479-86.
12	Zhang H, Ma JH, Wang J, et al. Statistical image reconstruction for low-dose CT using nonlocal means-based regularization. Part II: an adaptive approach[J]. Comput Med Imaging Graph, 2015, 43: 26-35.
13	Zhang H, Ma JH, Wang J, et al. Statistical image reconstruction for low-dose CT using nonlocal means-based regularization. Part II: an adaptive approach[J]. Comput Med Imaging Graph, 2015, 43: 26-35.
14	Niu SZ, Liu H, Zhang MZ, et al. Iterative reconstruction for low-dose cerebral perfusion computed tomography using prior image induced diffusion tensor[J]. Phys Med Biol, 2021, 66(11): 1150-24.
15	Wang J, Li TF, Lu HB, et al. Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography[J]. IEEE Trans Med Imaging, 2006, 25(10): 1272-83.
16	Wang J, Lu HB, Wen JH, et al. Multiscale penalized weighted least-squares sinogram restoration for low-dose X-ray computed tomography[J]. IEEE Trans Biomed Eng, 2008, 55(3): 1022-31.
17	张心如, 周先春, 汪志飞, et al. 基于改进各向异性扩散的图像去噪算法 [J]. 电子测量技术, 2022, 45(17): 113-9.
18	Guo YH, Cheng HD. Image noise removal approach based on subpixel anisotropic diffusion[J]. J Electron Imaging, 2012, 21(3): 033026-1.
19	Zhang HJ, Zeng D, Lin JH, et al. Iterative reconstruction for dual energy CT with an average image-induced nonlocal means regularization[J]. Phys Med Biol, 2017, 62(13): 5556-74.
20	Huang J, Zhang YW, Ma JH, et al. Iterative image reconstruction for sparse-view CT using normal-dose image induced total variation prior[J]. PLoS One, 2013, 8(11): e79709.
21	Zhang HW, Zhang PC, Cheng WT, et al. Learnable PM diffusion coefficients and reformative coordinate attention network for low dose CT denoising[J]. Phys Med Biol, 2023, 68(24): Physicsinmedi-cineandbiologyvol.68, 2410.1088/1361-6560/aced33.11Dec.2023, .
22	Wang J, Guan HQ, Solberg T. Inverse determination of the penalty parameter in penalized weighted least-squares algorithm for noise reduction of low-dose CBCT[J]. Med Phys, 2011, 38(7): 4066-72.
23	Liu Y, Ma JH, Fan Y, et al. Adaptive-weighted total variation minimization for sparse data toward low-dose X-ray computed tomography image reconstruction[J]. Phys Med Biol, 2012, 57(23): 7923-56.
24	牛善洲, 刘宏, 朱赟, et al. 基于广义惩罚加权最小二乘的低剂量CT重建方法 [J]. 数值计算与计算机应用, 2021, 42(3): 289-302. DOI: 10.12288/szjs.s2020-0671
25	牛善洲, 张梦真, 邱洋, et al. 基于全广义变分约束加权最小二乘的低剂量计算机断层重建方法[J]. 激光与光电子学进展, 2023, 60(4): 3788/LOP212853. DOI: 10.3788/LOP212853
26	Zhang H, Capaldi D, Zeng D, et al. Prior-image-based CT reconstruction using attenuation-mismatched priors[J]. Phys Med Biol, 2021, 66(6): 064007.
27	Zhang L, Zhang L, Mou XQ, et al. FSIM: a feature similarity index for image quality assessment[J]. IEEE Trans Image Process, 2011, 20(8): 2378-86.
28	Tsai MY, Liang HL, Chuo CC, et al. A novel protocol for abdominal low-dose CT scans adapted with a model-based iterative reconstruction method[J]. J Xray Sci Technol, 2023, 31(3): 453-61.
29	Wilson DO. Is there more to low dose CT scans[J]? Chest, 2022, 161(4): 880-1.
30	Sun T, Sun NB, Wang J, et al. Iterative CBCT reconstruction using Hessian penalty[J]. Phys Med Biol, 2015, 60(5): 1965-87.
31	Niu S, Bian Z, Zeng D, et al. Total image constrained diffusion tensor for spectral computed tomography reconstruction[J]. Applied Mathematical Modelling, 2019, 68: 487-508.
32	Niu SZ, Zhang MZ, Qiu Y, et al. Evaluation of low-dose computed tomography reconstruction using spatial-radon domain total generalized variation regularization[J]. Phys Med Biol, 2024, 69(10): 1050-05.

[1]	NIU Shanzhou, TANG Shizhou, HUANG Shuyan, LIANG Lijing, LI Shuo, LIU Hanming. Low-dose CT reconstruction based on high-dimensional partial differential equation projection recovery [J]. Journal of Southern Medical University, 2024, 44(4): 682-688.
[2]	CHEN Shixuan, ZENG Dong, BIAN Zhaoying, MA Jianhua. A low-dose CT reconstruction algorithm across different scanners based on federated feature learning [J]. Journal of Southern Medical University, 2024, 44(2): 333-343.
[3]	WANG Hailong, LIN Guoqin, DUAN Xiaoman, QI Mengke, WU Wangjiang, MA Jianhui, XU Yuan. A method for sensitivity analysis of deviation factor for geometric correction of cone-beam CT system [J]. Journal of Southern Medical University, 2023, 43(7): 1233-1240.
[4]	SI Wenbin, FENG Yanqiu. A multi-channel input convolutional neural network for artifact reduction in quantitative susceptibility mapping [J]. Journal of Southern Medical University, 2022, 42(12): 1799-1806.
[5]	HUANG Jinhong, ZHOU Genjiao, YU Zefeng, HU Wenyu. Deep parallel MRI reconstruction based on a complex-valued loss function [J]. Journal of Southern Medical University, 2022, 42(12): 1755-1764.
[6]	. Super-resolution construction of intravascular ultrasound images using generative adversarial networks [J]. Journal of Southern Medical University, 2019, 39(01): 82-.
[7]	. Design and optimization of a cone-beam CT system for extremity imaging [J]. Journal of Southern Medical University, 2018, 38(11): 1331-.
[8]	. Super-resolution reconstruction of lung 4D-CT images based on fast sub-pixel motion estimation [J]. Journal of Southern Medical University, 2015, 35(07): 1034-.
[9]	. Cone beam CT image iterative reconstruction based on Split-Bregman method [J]. Journal of Southern Medical University, 2014, 34(06): 783-.