An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson–Gaussian noise

Zhang, Shu; Xia, Youshen; Zou, Changzhong

doi:10.1016/j.ijleo.2019.04.005

Cited by 10 publications

(4 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…While SIR methods can control noise in the reconstructed images, the interactions among the regularization, system models, and statistical weighting (which are explicit in PWLS and implicit in PL) can cause the reconstructed images to have nonuniform spatial resolution and noise properties, even for idealized shift‐invariant imaging systems . Many regularizations have been designed for SIR to achieve approximately uniform resolution or uniform noise properties for emission tomography and x‐ray CT. These regularizations take the following form:

U false(boldμ false) = \underset{j}{false\sum} \underset{m \in W_{j}}{false\sum} w_{italicjm} {\overset{false^}{κ}}_{j} {\overset{false^}{κ}}_{m} ϕ (μ_{j} - μ_{m})

where

{\overset{false^}{k}}_{j}

is termed “pre‐tuned spatial strength” for x‐ray CT, controlling local spatial resolution and noise in the reconstructed image.…”

Section: Regularization Strategiesmentioning

confidence: 99%

Regularization strategies in statistical image reconstruction of low‐dose x‐ray CT: A review

et al. 2018

View full text Add to dashboard Cite

Statistical image reconstruction (SIR) methods have shown potential to substantially improve the image quality of low-dose x-ray computed tomography (CT) as compared to the conventional filtered back-projection (FBP) method. According to the maximum a posteriori (MAP) estimation, the SIR methods are typically formulated by an objective function consisting of two terms: (a) a data-fidelity term that models imaging geometry and physical detection processes in projection data acquisition, and (b) a regularization term that reflects prior knowledge or expectations of the characteristics of the to-be-reconstructed image. SIR desires accurate system modeling of data acquisition, while the regularization term also has a strong influence on the quality of reconstructed images. A variety of regularization strategies have been proposed for SIR in the past decades, based on different assumptions, models, and prior knowledge. In this paper, we review the conceptual and mathematical bases of these regularization strategies and briefly illustrate their efficacies in SIR of low-dose CT.

show abstract

U false(boldμ false) = \underset{j}{false\sum} \underset{m \in W_{j}}{false\sum} w_{italicjm} {\overset{false^}{κ}}_{j} {\overset{false^}{κ}}_{m} ϕ (μ_{j} - μ_{m})

where

{\overset{false^}{k}}_{j}

is termed “pre‐tuned spatial strength” for x‐ray CT, controlling local spatial resolution and noise in the reconstructed image.…”

Section: Regularization Strategiesmentioning

confidence: 99%

Regularization strategies in statistical image reconstruction of low‐dose x‐ray CT: A review

et al. 2018

View full text Add to dashboard Cite

show abstract

“…Nowadays, regularization methods and least squares methods are applied in CT reconstruction [34,35]. The most popular regularization method is Tikhonov regularization [25,36], which adds a constant to the eigenvalue to improve the stability of the matrix.…”

Section: Introductionmentioning

confidence: 99%

Optimized CT-TDLAS reconstruction performance evaluation of least squares with the polynomial-fitting method

Zhou

Cui

et al. 2022

Front. Phys.

View full text Add to dashboard Cite

Computed tomography-tunable diode laser absorption spectroscopy (CT-TDLAS) has been widely used in the diagnosis of the combustion flow field. Several optimized CT reconstruction algorithms such as iteration methods, transformation methods, and nonlinear least squares were applied. Considering the industrial application background, the performances of algebraic iteration reconstruction with the simultaneous algebra reconstruction technique (SART), Tikhonov regularization, and least squares with the polynomial fitting method were discussed in this study. For the mentioned algorithm, identical simulated reconstruction parameters that contained 32-path laser structures, assumed temperature distribution, and absorption databases were adopted to evaluate the reconstruction performance including accuracy, efficiency, and measurement of environment applicability. In this study, different CT reconstruction algorithms were also used to calculate the temperature distribution of the Bunsen burner flame. The different reconstruction results were compared with thermocouple detection data. With the theoretically simulated and experimental analysis, the least squares with the polynomial fitting technique has advantages in reconstruction accuracy, calculation efficiency, and laser path applicability for the measurement condition. It will be helpful in enhancing CT-TDLAS technique development.

show abstract

“…Due to the low image quality, works have been performed on applying the reconstruction approach to generate CT images using sparse-angular sampling. To rebuild CT images from severely under-sampled data, compressed sensing (CS) [10][11][12][13][14] has been proposed. This approach formulates the reconstruction issue as a convex optimization problem to promote image sparsity.…”

Section: Introductionmentioning

confidence: 99%

Complex artefact suppression for sparse reconstruction based on compensation approach in X‐ray computed tomography

Yang

Zhang

Huang

et al. 2022

IET Image Processing

View full text Add to dashboard Cite

To provide high-quality imaging by decreasing the sparse view imaging artefact from the computed tomography (CT) images, this study addresses a new artefact suppression technique for sparse data for both single material and multi-material objects that have diverse materials. It begins with a pre-reconstructed image and a network of target patches that have been trained beforehand and then uses the forward projection (FP) approach to resolve the structural mutation brought on by sparse-view projection. As an edge-preserving operator to commit to the forward operator for sinogram correction, the bilateral filter was used. Both simulated and actual data have been gathered and evaluated in experiments. The suggested forward operator and normalized compensation (FONC) method produce results that have far smaller artefact and errors than those of more traditional techniques. For simulation # blade, the Normalized Mean Square Distance (NMSD) of the proposed method was reduced by 8.94%, Structural Similarity Index (SSIM) and the Universal Quality Index (UQI) were increased by 78.17% and 80.49%, respectively, which also demonstrate better uniformity of the results to practical data # pan, where the root mean squared error was reduced by 13.93%. SSIM and UQI were increased by 25.66% and 37.02%, respectively. The results conclusively show that the planned strategies are successful in eliminating artefact for irregular objects.

show abstract

An adaptive regularization method for low-dose CT reconstruction from CT transmission data in Poisson–Gaussian noise

Cited by 10 publications

References 50 publications

Regularization strategies in statistical image reconstruction of low‐dose x‐ray CT: A review

Regularization strategies in statistical image reconstruction of low‐dose x‐ray CT: A review

Optimized CT-TDLAS reconstruction performance evaluation of least squares with the polynomial-fitting method

Complex artefact suppression for sparse reconstruction based on compensation approach in X‐ray computed tomography

Contact Info

Product

Resources

About