A weighted rebinned backprojection‐filtration algorithm from partially beam‐blocked data for a single‐scan cone‐beam CT with hybrid type scatter correction

Abstract

Purpose Scatter contamination constitutes a dominant source of degradation of image quality in cone‐beam computed tomography (CBCT). We have recently developed an analytic image reconstruction method with a scatter correction capability from the partially blocked cone‐beam data out of a single scan. Despite its easy implementation and its computational efficiency, the developed method may result in additional image artifacts for a large cone angle geometry due to data inconsistency. To improve the image quality at a large cone angle, we propose a weighted rebinned backprojection‐filtration (wrBPF) algorithm in conjunction with a hybrid type scatter correction approach. Methods The proposed method uses a beam‐blocker array that provides partial data for scatter correction and image reconstruction and that only blocks the beam within a limited cone angle. This design allows a chance to keep the image quality at larger cone angles by use of data redundancy since the projection data corresponding to larger cone angles are not blocked. However, the scatter correction would not be straightforward. In order to correct for the scatter in the projections at larger cone angles, we propose a novel scatter correction method combining a measurement‐based and a convolution‐based method. We first estimated the scatter signal using a measurement‐based method in the partially beam‐blocked regions, and then optimized the fitting parameters of a convolution‐kernel that can be used for scatter correction in the projections at larger cone angles. For image reconstruction, we developed a wrBPF with butterfly filtering. We have conducted an experimental study to validate the proposed algorithm for image reconstruction and scatter correction. Results The experimental results revealed that the developed reconstruction method makes full use of the benefits of partial beam‐blocking for scatter correction and image reconstruction and at the same time enhances image quality at larger cone angles by use of an optimized convolution‐based scatter correction. Conclusions The proposed method that enjoys the advantages of both measurement‐based and convolution‐based methods for scatter correction has successfully demonstrated its capability of reconstructing accurate images out of a single scan in circular CBCT.

A practical and robust method for beam blocker-based cone beam CT scatter correction

“…Both scan time and patient dose are increased in the two-scan method, and the scatter correction performance may also be affected by the patient motion between two scans. To solve the problems, the moving blocker and half blocker methods are proposed to obtain scatter-corrected CBCT in a single scan by making full use of the property of signal redundancy (Niu and Zhu 2011, Lee et al 2012, Liang et al 2019, Min et al 2019. Since the primary signal loss is irreversible and can result in severe artifacts in the images, the data acquisition in the moving blocker method is a complicated process and should be implemented very cautiously (Niu and Zhu 2011).…”

A scatter correction method of CBCT via CycleGAN and forward projection algorithm

“…Scatter correction technology for CBCT has been widely study, those methods can be mainly divided three categories: software-based, [1][2][3][4] hardware-based 5,6 and software-hardware hybrid methods. [7][8][9][10][11] In earlier study, many hardware-based methods been proposed, such as anti-scatter grid method, air gap method and so on. Software-based mainly contains Monte Carlo, deep learning and SKS methods.…”

Planning CT‐guided robust and fast cone‐beam CT scatter correction using a local filtration technique

“…To improve CBCT image accuracy, various scatter estimation methods are proposed and could be divided into two categories: hardware-based [13][14][15][16][17][18][19][20] and softwarebased [21][22][23][24][25][26][27][28] methods. The former usually applies partial beam blocker [13][14][15][16][17] or primary modulator [18][19][20] to measure scatter signals on the flat-panel detector, while the latter exploits scatter-kernel modeling 21,22 and physical simulation. [23][24][25][26] All these methods obtain scatter estimation via either direct measurement or theoretical calculation.…”

“…To improve CBCT image accuracy, various scatter estimation methods are proposed and could be divided into two categories: hardware‐based 13–20 and software‐based 21–28 methods. The former usually applies partial beam blocker 13–17 or primary modulator 18–20 to measure scatter signals on the flat‐panel detector, while the latter exploits scatter‐kernel modeling 21,22 and physical simulation 23–26 .…”