Robust Low-Dose CT Sinogram Preprocessing via Exploiting Noise-Generating Mechanism.

Area of Science:

• Medical Imaging
• Image Processing
• Radiology

Background:

• Low-dose computed tomography (CT) images suffer from severe degradation due to physical factors and noise.
• Existing sinogram recovery methods often use general image priors that do not fully capture sinogram characteristics.

Purpose of the Study:

• To develop an improved sinogram preprocessing method for low-dose CT image recovery.
• To address the limitations of traditional methods by incorporating accurate statistical properties and a novel image prior.

Main Methods:

• Formulated low-dose CT sinogram preprocessing as a maximum a posteriori (MAP) estimation problem.
• Incorporated statistical properties of X-ray photon and electronic noise.
• Designed a new prior formulation to encode piecewise-linear configurations in sinograms, improving upon traditional TV prior terms.
• Developed an efficient alternating direction method of multipliers (ADMM) algorithm to solve the MAP framework.

Main Results:

• The proposed model's likelihood/loss and prior/regularization terms are more accurate and aligned with the sinogram generation mechanism.
• Experiments on simulated and real low-dose CT data showed superior performance compared to previous methods.
• Both visual inspection and quantitative evaluations confirmed the effectiveness of the proposed method.

Conclusions:

• The novel MAP estimation framework with an improved prior significantly enhances low-dose CT sinogram recovery.
• The proposed method offers a more statistically robust and accurate approach to preprocessing low-dose CT data.
• This advancement holds potential for improving diagnostic accuracy in low-dose CT imaging.