Miran Park

Data consistency-driven scatter kernel optimization for x-ray cone-beam CT

Accurate and efficient scatter correction is essential for acquisition of high-quality x-ray cone-beam CT (CBCT) images for various applications. This study was conducted to demonstrate the feasibility of using the data consistency condition (DCC) as a criterion for scatter kernel optimization in scatter deconvolution methods in CBCT. As in CBCT, data consistency in the mid-plane is primarily challenged by scatter, we utilized data consistency to confirm the degree of scatter correction and to steer the update in iterative kernel optimization. By means of the parallel-beam DCC via fan-parallel rebinning, we iteratively optimized the scatter kernel parameters, using a particle swarm optimization algorithm for its computational efficiency and excellent convergence. The proposed method was validated by a simulation study using the XCAT numerical phantom and also by experimental studies using the ACS head phantom and the pelvic part of the Rando phantom. The results showed that the proposed method can effectively improve the accuracy of deconvolution-based scatter correction. Quantitative assessments of image quality parameters such as contrast and structure similarity (SSIM) revealed that the optimally selected scatter kernel improves the contrast of scatter-free images by up to 99.5%, 94.4%, and 84.4%, and of the SSIM in an XCAT study, an ACS head phantom study, and a pelvis phantom study by up to 96.7%, 90.5%, and 87.8%, respectively. The proposed method can achieve accurate and efficient scatter correction from a single cone-beam scan without need of any auxiliary hardware or additional experimentation.

Sparse-view computed laminography with a spherical sinusoidal scan for nondestructive testing

X-ray computed laminography is widely used in nondestructive testing of relatively flat objects using an oblique scanning configuration for data acquisition. In this work, a new scanning scheme is proposed in conjunction with the compressive-sensing-based image reconstruction for reducing imaging radiation dose and scanning time. We performed a numerical study comparing image qualities acquired by various scanning configurations that are practically implementable: single-arc, double-arc, oblique, and spherical-sinusoidal trajectories. A compressive-sensing-inspired total-variation (TV) minimization algorithm was used to reconstruct the images from the data acquired at only 40 projection views in those trajectories. It was successfully demonstrated that the proposed scanning scheme outperforms the others in terms of image contrast and spatial resolution, although the oblique scanning scheme showed a comparable resolution property. We believe that the proposed scanning method may provide a solution to fast and low-dose nondestructive testing of radiation-sensitive and highly integrated devices such as multilayer microelectronic circuit boards.

A Feasibility Study of Low-Dose Single-Scan Dual-Energy Cone-Beam CT in Many-View Under-Sampling Framework

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.

Scatter correction in CBCT with an offset detector through a deconvolution method using data consistency

Our earlier work has demonstrated that the data consistency condition can be used as a criterion for scatter kernel optimization in deconvolution methods in a full-fan mode cone-beam CT [1]. However, this scheme cannot be directly applied to CBCT system with an offset detector (half-fan mode) because of transverse data truncation in projections. In this study, we proposed a modified scheme of the scatter kernel optimization method that can be used in a half-fan mode cone-beam CT, and have successfully shown its feasibility. Using the first-reconstructed volume image from half-fan projection data, we acquired full-fan projection data by forward projection synthesis. The synthesized full-fan projections were partly used to fill the truncated regions in the half-fan data. By doing so, we were able to utilize the existing data consistency-driven scatter kernel optimization method. The proposed method was validated by a simulation study using the XCAT numerical phantom and also by an experimental study using the ACS head phantom.

X-ray inspection system with two flat panel detectors for extra-large object inspection

The medical imaging technologies and the industrial ones have benefitted from each other. For a baggage inspection for example, multiple-detector helical CT scanning technology has been deployed that was originally developed for a fast volumetric imaging of a patient in clinics. In a helical CT, the detector array and the x-ray source are rotating continuously in a gantry with the slip-ring technology, and the object is moving along the rotation axis. With the slip-ring technology, an elaborate x-ray cable and drum system can be eliminated but may increase the production cost and instability of the system.

Improving image quality of a mobile Cone-Beam CT by use of scatter and beam-hardening corrections

Cone-beam computed tomography (CBCT) imaging system has been widely used in various applications, and takes an important role for diagnosis of emergency patient. However large size of detector causes high scatter-to-primary ratios on CBCT projection data. Also, beam hardening artifacts in reconstructed CT images is resulted by the polychromatic spectrum of X-ray beam. These two phenomena lead to unwanted results of reconstructed image such as cupping artifact, worsen soft tissue contrast, and lower CT grey value. This paper studies the result of investigation into the effect of the scatter kernel superposition (SKS) method and polynomial beam hardening correction on the measurement soft tissue contrast of head phantom for reconstructed data from half-fan CBCT system. The measurement imaging system was PHION, which is mobile CT system of NanoFocusRay Co., Ltd, and PH-3 angiographic CT head phantom ACS was used for experiments at 120 kVp and 15 mAs. The data was processed by scatter correction and beam hardening correction which is a linearization method based on polynomial curves of second order. The image was reconstructed using a filtered back projection algorithm. After scatter and beam hardening correction, contrast to noise ratio of reconstructed images is increased and soft tissues are seen more clearly. Image with scatter correction and beam hardening correction has clear boundary between skull and soft tissues, but only scatter corrected image has unclear boundary.

Feasibility study on multiple fan-beam data acquisition for low-dose helical CT

In computed tomography (CT) imaging, radiation dose delivered to the patient is one of the major concerns. Many CT developers and researchers have been making efforts to reduce radiation dose. Sparse-view CT takes projections at sparser view-angles and provides a viable option to reducing radiation dose. However, a fast power switching of an x-ray tube, which is needed for the sparse-view sampling, can be challenging in many CT systems. We have recently proposed a novel alternative approach to sparse-view circular CT that can be readily incorporated in the existing CT systems. Instead of switching the x-ray tube power, we proposed to use a multi-slit collimator placed between the x-ray source and the patient to partially block the x-ray beam thereby reducing the radiation. In this study, we performed a simulation study based on numerically acquired projection data to demonstrate a feasibility of using a multi-slit collimator in a helical CT. The XCAT phantom was used and a numerical collimator has been made to apply on the projection data. Numerical multi-slit collimator was designed to have equal size of slit-openings and radio-opaque rectangular areas, and the length dimension of the slits is perpendicular to the rotation axis. For image reconstruction, we used a total-variation minimization (TV) algorithm which has shown its out-performance in many sparse-view CT applications. We demonstrated that the proposed multiple fan-beam helical CT can provide a useful low-dose scanning option.

Bunched sparse-view CT using a moving multi-slit collimator

In computed tomography (CT), useful anatomical information can be obtained and used for various clinical purposes. However, reduction of radiation dose is one of the important issues. The aim of our work is to decrease radiation dose using an oscillating multi-slit collimator while preserving image quality. The collimator was set to mask data perpendicular to the rotation axis, and it was driven to oscillate parallel to the rotation axis. The oscillating motion was programmed to have a sinusoidal motion. Varying sizes of slit-openings were simulated in this work: half, one-fourth, and one-eighth openings. We used a total variation (TV) minimization algorithm to reconstruct the images from these bunched sparse-view data. Universal quality Index (UQI) was calculated to compare the image similarities to the reference image. This work successfully demonstrated a feasibility of the proposed method for low-dose CT. This method may be implemented in the existing CT systems without much technical difficulties.