Abstract
In spectral CT imaging, the coefficient image of the basis material obtained by the material decomposition technique can estimate the tissue composition, and its accuracy directly affects the disease diagnosis. Although the precision of material decomposition is increased by employing convolutional neural networks (CNN), extracting the non-local features from the CT image is restricted using the traditional CNN convolution operator. A graph model built by multi-scale non-local self-similar patterns is introduced into multi-material decomposition (MMD). We proposed a novel MMD method based on graph edge–conditioned convolution U-net (GECCU-net) to enhance material image quality. The GECCU-net focuses on developing a multi-scale encoder. At the network coding stage, three paths are applied to capture comprehensive image features. The local and non-local feature aggregation (LNFA) blocks are designed to integrate the local and non-local features from different paths. The graph edge–conditioned convolution based on non-Euclidean space excavates the non-local features. A hybrid loss function is defined to accommodate multi-scale input images and avoid over-smoothing of results. The proposed network is compared quantitatively with base CNN models on the simulated and real datasets. The material images generated by GECCU-net have less noise and artifacts while retaining more information on tissue. The Structural SIMilarity (SSIM) of the obtained abdomen and chest water maps reaches 0.9976 and 0.9990, respectively, and the RMSE reduces to 0.1218 and 0.4903 g/cm3. The proposed method can improve MMD performance and has potential applications.
Graphical Abstract
Similar content being viewed by others
References
Bhayana R, Parakh A, Kambadakone A (2020) Material decomposition with dual-and multi-energy computed tomography. MRS Commun 10:558–565. https://doi.org/10.1557/mrc.2020.86
Franco PN, Spasiano CM, Maino C, De Ponti E, Ragusi M, Giandola T, Terrani S, Peroni M, Corso R, Ippolito D (2023) Principles and applications of dual-layer spectral CT in gastrointestinal imaging. Diagnostics 13:1740. https://doi.org/10.3390/diagnostics13101740
Wang S, Cai A, Wu W, Zhang T, Liu F, Yu H (2023) IMD-MTFC: image-domain material decomposition via material-image tensor factorization and clustering for spectral CT. IEEE Trans Radiation Plasma Med Sci 7:382–393. https://doi.org/10.1109/TRPMS.2023.3234613
Fang W, Wu DF, Kim K, Kalra MK, Singh R, Li L, Li QZ (2021) Iterative material decomposition for spectral CT using self-supervised Noise2Noise prior. Phys Med Biol 66. https://doi.org/10.1088/1361-6560/ac0afd
Geng MF, Tian ZF, Jiang Z, You YF, Feng XM, Xia Y, Yang K, Ren QS, Meng XX, Maier A et al (2021) PMS-GAN: parallel multi-stream generative adversarial network for multi-material decomposition in spectral computed tomography. IEEE Trans Med Imaging 40:571–584. https://doi.org/10.1109/TMI.2020.3031617
Xue Y, Jiang YK, Yang CL, Lyu QH, Wang J, Luo C, Zhang LH, Desrosiers C, Feng K, Sun XN et al (2019) Accurate multi-material decomposition in dual-energy CT: a phantom study. IEEE Transactions Comput Imaging 5:515–529. https://doi.org/10.1109/TCI.2019.2909192
Barber RF, Sidky EY, Schmidt TG, Pan XC (2016) An algorithm for constrained one-step inversion of spectral CT data. Phys Med Biol 61:3784–3818. https://doi.org/10.1088/0031-9155/61/10/3784
Feng M, Ji X, Zhang R, Treb K, Dingle AM, Li K (2021) An experimental method to correct low-frequency concentric artifacts in photon counting CT. Phys Med Biol 66:175011. https://doi.org/10.1088/1361-6560/ac1833
Kim B, Shim H, Baek J (2022) A streak artifact reduction algorithm in sparse-view CT using a self-supervised neural representation. Med Phys 49:7497–7515. https://doi.org/10.1002/mp.15885
He Y, Zeng L, Xu Q, Wang Z, Yu H, Shen Z, Yang Z, Zhou R (2023) Spectral CT reconstruction via low-rank representation and structure preserving regularization. Phys Med Biol 68:025011. https://doi.org/10.1109/TMI.2020.2983414
Ren L, Mccollough CH, Yu L (2018) Three-material decomposition in multi-energy CT: impact of prior information on noise and bias. In:SPIE, p 363–368. https://doi.org/10.1117/12.2294953
Tao SZ, Rajendran K, Mccollough CH, Leng S (2018) Material decomposition with prior knowledge aware iterative denoising (MD-PKAID). Phys Med Biol 63. https://doi.org/10.1088/1361-6560/aadc90
Zavala-Mondragon LA, Engel KJ, Menser B, Ruijters D, Van Der Sommen F (2021) Iterative reconstruction anti-correlated ROF model for noise reduction in dual-energy CBCT imaging. In: SPIE, p 661–670. https://doi.org/10.1117/12.2579500
Wu WW, Yu HJ, Chen PJ, Luo FL, Liu FL, Wang Q, Zhu YN, Zhang YB, Feng J, Yu HY (2020) Dictionary learning based image-domain material decomposition for spectral CT. Phys Med Biol 65. https://doi.org/10.1088/1361-6560/aba7ce
Niu SZ, Zhang Y, Zhong YC, Liu GL, Lu SH, Zhang XL, Hu SZ, Wang TH, Yu GH, Wang J (2018) Iterative reconstruction for photon-counting CT using prior image constrained total generalized variation. Comput Biol Med 103:167–182. https://doi.org/10.1016/j.compbiomed.2018.10.022
Yang QS, Yan PK, Zhang YB, Yu HY, Shi YY, Mou XQ, Kalra MK, Zhang Y, Sun L, Wang G (2018) Low-dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss. IEEE Trans Med Imaging 37:1348–1357. https://doi.org/10.1109/TMI.2018.2827462
Shi ZF, Wang N, Kong FN, Cao HS, Cao QJ (2022) A semi-supervised learning method of latent features based on convolutional neural networks for CT metal artifact reduction. Med Phys 49:3845–3859. https://doi.org/10.1002/mp.15633
Yu LQ, Zhang ZC, Li XM, Xing L (2021) Deep sinogram completion with image prior for metal artifact reduction in CT images. IEEE Trans Med Imaging 40:228–238. https://doi.org/10.1109/TMI.2020.3025064
You CY, Li G, Zhang Y, Zhang XL, Shan HM, Li MZ, Ju SH, Zhao Z, Zhang ZY, Cong WX et al (2020) CT super-resolution GAN constrained by the identical, residual, and cycle learning ensemble (GAN-CIRCLE). IEEE Trans Med Imaging 39:188–203. https://doi.org/10.1109/TMI.2019.2922960
Zhang SN, Zhao ZQ, Qiu L, Liang D, Wang K, Xu J, Zhao J, Sun JQ (2023) Automatic vertebral fracture and three-column injury diagnosis with fracture visualization by a multi-scale attention-guided network. Med Biol Eng Compu 61:1661–1674. https://doi.org/10.1007/s11517-023-02805-2
Apostolopoulos ID, Pintelas EG, Livieris IE, Apostolopoulos DJ, Papathanasiou ND, Pintelas PE, Panayiotakis GS (2021) Automatic classification of solitary pulmonary nodules in PET/CT imaging employing transfer learning techniques. Med Biol Eng Compu 59:1299–1310. https://doi.org/10.1007/s11517-021-02378-y
An MJ, Li JH, Xu XY, Schoepf UJ, Savage RH, Cao KL, Song Q, Wang ZY, Liu Z, Li YW et al (2023) A deep learning-based fully automatic and clinical-ready framework for regional myocardial segmentation and myocardial ischemia evaluation. Med Biol Eng Compu. https://doi.org/10.1007/s11517-023-02798-y
Hong ZF, Chen MZ, Hu WJ, Yan SY, Qu AP, Chen LN, Chen JX (2023) Dual encoder network with transformer-CNN for multi-organ segmentation. Med Biol Eng Compu 61:661–671. https://doi.org/10.1007/s11517-022-02723-9
Kawahara D, Saito A, Ozawa S, Nagata Y (2021) Image synthesis with deep convolutional generative adversarial networks for material decomposition in dual-energy CT from a kilovoltage CT. Comput Biol Med 128. https://doi.org/10.1016/j.compbiomed.2020.104111
Wang GS, Liu Z, Huang ZY, Zhang N, Luo HH, Liu LJ, Shen H, Che CW, Niu TY, Liang D et al. (2022) Improved GAN: using a transformer module generator approach for material decomposition. Comput Biol Med 149. https://doi.org/10.1016/j.compbiomed.2022.105952
Clark DP, Holbrook M, Badea CT (2018) Multi-energy CT decomposition using convolutional neural networks. In: Medical Imaging 2018: Phys Med Imaging https://doi.org/10.1117/12.2293728
Wu XC, He P, Long ZR, Guo XD, Chen MY, Ren XZ, Chen PJ, Deng LZ, An K, Li PC et al (2019) Multi-material decomposition of spectral CT images via Fully Convolutional DenseNets. J X-Ray Sci Technol 27:461–471. https://doi.org/10.3233/XST-190500
Gong H, Tao SZ, Rajendran K, Zhou W, Mccollough CH, Leng S (2020) Deep-learning-based direct inversion for material decomposition. Med Phys 47:6294–6309. https://doi.org/10.1002/mp.14523
Salehjahromi M, Zhang Y, Yu H (2017) A spectral CT denoising algorithm based on weighted block matching 3D filtering. In:SPIE, p 65–76. https://doi.org/10.1117/12.2273213
Zhang Y, Salehjahromi M, Yu H (2019) Tensor decomposition and non-local means based spectral CT image denoising. J Xray Sci Technol 27:397–416. https://doi.org/10.3233/XST-180413
Bruna J, Zaremba W, Szlam A, Lecun Y (2013) Spectral networks and locally connected networks on graphs. arXiv preprint arXiv:1312.6203. https://doi.org/10.48550/arXiv.1312.6203
Ding Y, Zhang Z, Zhao X, Hong D, Cai W, Yu C, Yang N, Cai W (2022) Multi-feature fusion: graph neural network and CNN combining for hyperspectral image classification. Neurocomputing 501:246–257. https://doi.org/10.1016/j.neucom.2022.06.031
Xuan P, Wu X, Cui H, Jin Q, Wang L, Zhang T, Nakaguchi T, Duh HBL (2023) Multi-scale random walk driven adaptive graph neural network with dual-head neighboring node attention for CT segmentation. Appl Soft Comput 133:109905. https://doi.org/10.1016/j.asoc.2022.109905
Gürler Z, Gharsallaoui MA, Rekik I, Alzheimer’s Dis Neuroimaging I (2023) Template-based graph registration network for boosting the diagnosis of brain connectivity disorders. Computer Med Imaging Graph 103. https://doi.org/10.1016/j.compmedimag.2022.102140
Zhou Y, Zheng HX, Huang X, Hao SF, Li DA, Zhao JM (2022) Graph neural networks: taxonomy, advances, and trends. ACM Trans Intell Syst Technol 13. https://doi.org/10.1145/3495161
He YJ, Zhao H, Wong STC (2021) Deep learning powers cancer diagnosis in digital pathology. Computer Med Imaging Graph 88. https://doi.org/10.1016/j.compmedimag.2020.101820
Valsesia D, Fracastoro G, Magli E (2020) Deep graph-convolutional image denoising. IEEE Trans Image Process 29:8226–8237. https://doi.org/10.1109/TIP.2020.3013166
Chen K, Pu X, Ren Y, Qiu H, Li H, Sun J (2020) Low-dose CT image blind denoising with graph convolutional networks. In: Kwok JT, Chan JH, King I (eds) Yang H, Pasupa K, Leung AC-S. Neural information processing. Springer International Publishing, Cham, pp 423–435
Shi ZF, Li JZ, Li HL, Hu QX, Cao QJ (2019) A virtual monochromatic imaging method for spectral CT based on Wasserstein generative adversarial network with a hybrid loss. IEEE Access 7:110992–111011. https://doi.org/10.1109/ACCESS.2019.2934508
Shi ZF, Li HL, Cao QJ, Wang ZQ, Cheng M (2021) A material decomposition method for dual-energy CT via dual interactive Wasserstein generative adversarial networks. Med Phys 48:2891–2905. https://doi.org/10.1002/mp.14828
Simonovsky M, Komodakis N, Ieee (2017) Dynamic edge-conditioned filters in convolutional neural networks on graphs. In: 30TH IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2017). p 29–38
Chen GH, Tang J, Leng SH (2008) Prior image constrained compressed sensing (PICCS): a method to accurately reconstruct dynamic CT images from highly undersampled projection data sets. Med Phys 35:660–663. https://doi.org/10.1118/1.2836423
Segars WP, Sturgeon G, Mendonca S, Grimes J, Tsui BMW (2010) 4D XCAT phantom for multimodality imaging research. Med Phys 37:4902–4915. https://doi.org/10.1118/1.3480985
Shi ZF, Yang HY, Cong WX, Wang G (2016) An edge-on charge-transfer design for energy-resolved x-ray detection. PHYSICS IN MEDICINE AND BIOLOGY 61:4183–4200. https://doi.org/10.1088/0031-9155/61/11/4183
Walsh MF, Nik SJ, Procz S, Pichotka M, Bell ST, Bateman CJ, Doesburg RMN, De Ruiter N, Chernoglazov AI, Panta RK et al. (2013) Spectral CT data acquisition with Medipix3.1. J Instrument 8. https://doi.org/10.1088/1748-0221/8/10/P10012
Chandra TB, Verma K (2020) Analysis of quantum noise-reducing filters on chest X-ray images: a review. Measurement 153:107426. https://doi.org/10.1016/j.measurement.2019.107426
Zhang L, Zhang L, Mou XQ, Zhang D (2011) FSIM: a feature similarity index for image quality assessment. IEEE Trans Image Process 20:2378–2386. https://doi.org/10.1109/TIP.2011.2109730
Kingma D P, Ba J (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980. https://doi.org/10.48550/arXiv.1412.6980
Acknowledgements
The authors would like to thank Prof. Segars from Duke University for providing the phantom datasets, and we are also grateful to Prof. Anthony Butler and Hannah Prebble from MARS Bioimaging Ltd. for sharing the mouse datasets.
Funding
This work was funded by the National Natural Science Foundation of China through grant 62071326.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Shi, Z., Kong, F., Cheng, M. et al. Multi-energy CT material decomposition using graph model improved CNN. Med Biol Eng Comput 62, 1213–1228 (2024). https://doi.org/10.1007/s11517-023-02986-w
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11517-023-02986-w