Noise level and no-reference-based evaluations for gamma-ray image using non-local means algorithm with CZT photon...

Area of Science:

• Medical Imaging
• Nuclear Physics
• Materials Science

Background:

• Thin cadmium zinc telluride (CZT) photon-counting semiconductor detectors (PCSDs) offer potential for high-resolution gamma-ray imaging.
• Optimizing image quality with thin detectors presents challenges in maintaining spatial resolution and signal-to-noise ratio.
• Existing imaging techniques may not fully leverage the capabilities of thin CZT PCSDs.

Purpose of the Study:

• To propose and evaluate an algorithm for maximizing gamma-ray image quality using thin CZT PCSDs.
• To investigate the impact of a pixel-matched parallel-hole collimator on spatial resolution.
• To assess the effectiveness of the non-local mean (NLM) noise reduction algorithm in improving image quality.

Main Methods:

• Modeling of a pixel-matched parallel-hole collimator for gamma-ray imaging.
• Application of the non-local mean (NLM) noise reduction algorithm using the Geant4 Application for Tomographic Emission platform.
• Quantitative evaluation of image quality metrics, including spatial resolution (Full Width at Half Maximum), contrast-to-noise ratio, and coefficient of variation.

Main Results:

• A thin CZT PCSD (1 mm) combined with the NLM algorithm achieved a spatial resolution of 0.957 mm (FWHM).
• Spatial resolution improved by approximately 40.89% compared to a 3-mm-thick PCSD under identical conditions.
• The NLM algorithm-enhanced system demonstrated contrast-to-noise ratio and coefficient of variation comparable to thicker detector systems.

Conclusions:

• The proposed NLM algorithm effectively enhances gamma-ray image quality from thin CZT PCSDs.
• This approach offers a cost-effective method for improving gamma-ray imaging technology.
• The study demonstrates the feasibility of achieving high spatial resolution with thinner, more economical detector configurations.