Scatter Correction Method for X-Ray CT Using Primary Modulation: Theory and Preliminary Results

Abstract

An X-ray system with a large area detector has high scatter-to-primary ratios (SPRs), which result in severe artifacts in reconstructed computed tomography (CT) images. A scatter correction algorithm is introduced that provides effective scatter correction but does not require additional patient exposure. The key hypothesis of the algorithm is that the high-frequency components of the X-ray spatial distribution do not result in strong high-frequency signals in the scatter. A calibration sheet with a checkerboard pattern of semitransparent blockers (a "primary modulator") is inserted between the X-ray source and the object. The primary distribution is partially modulated by a high-frequency function, while the scatter distribution still has dominant low-frequency components, based on the hypothesis. Filtering and demodulation techniques suffice to extract the low-frequency components of the primary and hence obtain the scatter estimation. The hypothesis was validated using Monte Carlo (MC) simulation, and the algorithm was evaluated by both MC simulations and physical experiments. Reconstructions of a software humanoid phantom suggested system parameters in the physical implementation and showed that the proposed method reduced the relative mean square error of the reconstructed image in the central region of interest from 74.2% to below 1%. In preliminary physical experiments on the standard evaluation phantom, this error was reduced from 31.8% to 2.3%, and it was also demonstrated that the algorithm has no noticeable impact on the resolution of the reconstructed image in spite of the filter-based approach. Although the proposed scatter correction technique was implemented for X-ray CT, it can also be used in other X-ray imaging applications, as long as a primary modulator can be inserted between the X-ray source and the imaged object.

Accelerated barrier optimization compressed sensing (ABOCS) reconstruction for cone-beam CT: Phantom studies

“…After the effective data correction algorithms already implemented on a CBCT system, including scatter and beamhardening corrections, [11][12][13][14][15][16][17] the residual nonstatistical errors on the projections are expected to be small as compared to the Poisson noise. In this paper, we assume that this condition is true.…”

“…Note that, the user-defined parameter ε is interpreted as the variance of the difference between the predicted and the raw projections. After effective data correction for scatter and beam-hardening effects, [11][12][13][14][15][16][17][18] most of the projection errors in CT scans are from Poisson statistics of the incident photons, except for very lowdose imaging cases. [19][20][21] ε can therefore be readily estimated from the measured projections.…”

Experimental Study of Intracranial Hematoma Detection with Flat Panel Detector C-Arm CT

“…The correction approach implemented here estimates the primary radiation from the projection data by using an iterative method involving a multiplicative correction step, 19 and further improvements by using iterative techniques, hardware approaches, or a combination of the 2 are possible. [20][21][22][23][24][25][26][27][28][29][30] The number of views taken over a partial-circle scan can also be a limiting factor for image quality if insufficient angular sampling (ie, too large an angle between views) occurs. The Feldkamp reconstruction algorithm 31 with a generalized Parker weighting scheme, 32 acquiring projections over at least 180°plus the fan angle of the x-ray beam, with 538 views spaced 0.4°apart, provided images that were free of view aliasing on the midplane.…”

“…The correction method used here applies a polynomial operation to restore a linear relationship between projection value and path length, assuming water-equivalent object conditions. 20 Although this software-based method reduces cupping effectively, it cannot completely remove the streaking and shading artifacts and other more sophisticated beam-hardening correction methods (eg, based on iterative modifications of x-ray projection images as described in Hsieh et al 39 ) could be used to further improve image quality.…”

An energy minimization method for the correction of cupping artifacts in cone‐beam CT

“…However, even using the fast algorithm, the formidably heavy computation load associated with Monte Carlo scatter estimation hinders its real‐life applications. The third category of methods estimates the contribution of scatter in the measured projection data and then subtracts it from the measured projection data, such as the beam‐stop array method, (10) the moving blocker method, (11) primary modula‐tion, 12 , 13 , 14 artifact‐suppressed dictionary learning method, 15 , 16 and others 17 , 18 . However, these types of correction methods add extra hardware to the CBCT system, which can add difficulty and complexity.…”