Bayesian statistical reconstruction for low-dose X-ray computed tomography using an adaptive-weighting nonlocal prior

Area of Science:

• Medical Imaging
• Radiology
• Computer Vision

Background:

• Reducing radiation dose in computed tomography (CT) is crucial for patient safety.
• Low-dose CT scans often suffer from significant quantum noise, degrading image quality.
• Traditional filtered back-projection (FBP) methods struggle with noise in low-dose scenarios.

Purpose of the Study:

• To improve the image quality of low-dose CT scans.
• To introduce a novel adaptive-weighting nonlocal (AW nonlocal) prior statistical reconstruction approach.
• To address the limitations of traditional local priors in preserving resolution and removing noise.

Main Methods:

• Developed a novel AW nonlocal prior for statistical reconstruction in CT.
• Utilized Bayesian statistical reconstruction methods.
• Exploited global image information adaptively and selectively.

Main Results:

• The proposed AW nonlocal prior effectively reduces noise in low-dose CT images.
• The method preserves image resolution while removing noise.
• Reconstructions demonstrated excellent performance in low-dose X-ray CT scans.

Conclusions:

• The AW nonlocal prior is a promising technique for enhancing low-dose CT image quality.
• This approach offers superior noise reduction and resolution preservation compared to traditional methods.
• The findings support the clinical utility of this novel reconstruction method for safer CT imaging.