IEEE Trans Med Imaging. 2017 Dec;36(12):2578-2587. doi: 10.1109/TMI.2017.2765760.
A single-scan dual-energy low-dose cone-beam CT (CBCT) imaging technique that exploits a multi-slit filter is proposed in this paper. The multi-slit filter installed between the x-ray source and the scanned object is reciprocated during a scan. The x-ray beams through the slits would generate relatively low-energy x-ray projection data, while the filtered beams would make high-energy projection data. An iterative image reconstruction algorithm that uses an adaptive-steepest-descent method to minimize image total-variation under the constraint of data fidelity was applied to reconstructing the image from the low-energy projection data. Since the high-energy projection data suffer from a substantially high noise level due to the beam filtration, we have developed a new algorithm that exploits the joint sparsity between the low- and high-energy CT images for image reconstruction of the high-energy CT image. The proposed image reconstruction algorithm uses a gradient magnitude image (GMI) of the low-energy CT image by regularizing the difference of GMIs of the low- and high-energy CT images to be minimized. The feasibility of the proposed technique has been demonstrated by the use of various phantoms in the experimental CBCT setup. Furthermore, based on the proposed dual-energy imaging, a material differentiation was performed and its potential utility has been shown. The proposed imaging technique produced promising results for its potential application to a low-dose single-scan dual-energy CBCT.
本文提出了一种利用多狭缝滤波器的单次扫描双能低剂量锥形束 CT（CBCT）成像技术。在扫描过程中，狭缝滤波器安装在 X 射线源和被扫描物体之间，并进行往复运动。通过狭缝的 X 射线束会产生相对低能的 X 射线投影数据，而过滤后的射线则会产生高能投影数据。我们应用了一种迭代图像重建算法，该算法使用自适应最速下降法，在数据保真度的约束下最小化图像全变差，从而从低能投影数据重建图像。由于高能投影数据由于射线过滤而受到相当高的噪声水平的影响，我们开发了一种新算法，该算法利用低能和高能 CT 图像之间的联合稀疏性来重建高能 CT 图像。所提出的图像重建算法使用低能 CT 图像的梯度幅度图像（GMI），通过正则化低能和高能 CT 图像的 GMIs 之间的差异来最小化。在实验性 CBCT 设备中使用各种体模证明了该技术的可行性。此外，基于所提出的双能成像，进行了材料区分，并展示了其潜在的应用。所提出的成像技术在低剂量单次扫描双能 CBCT 中的潜在应用方面取得了有前景的结果。
