Nonlinear polychromatic and noise artifacts in x-ray computed tomography images

Area of Science:

• Medical Imaging Physics
• Radiological Sciences
• Image Processing

Background:

• Computed tomography (CT) image quality is affected by noise, primarily quantum noise.
• Beam hardening is a known artifact in CT imaging, also related to X-ray attenuation.
• Understanding nonlinear effects is crucial for accurate CT image analysis.

Purpose of the Study:

• To theoretically study the nonlinear relationship between X-ray attenuation and image noise variance in CT.
• To investigate the similar effects of quantum noise and beam hardening on CT image quality.
• To analyze the spatial distribution and characteristics of noise in CT images.

Main Methods:

• Theoretical analysis of noise variance and beam hardening as functions of X-ray attenuation.
• Computer simulations to model and visualize the nonlinear dependencies.
• Numerical evaluation of noise distribution and its relation to projection data.

Main Results:

• Both quantum noise variance and beam hardening exhibit nonlinear dependencies on X-ray attenuation.
• These nonlinearities produce similar effects on overall CT image quality.
• The 2D noise variance distribution is a weighted superposition of backprojected powers of noiseless projection data.

Conclusions:

• Nonlinear phenomena significantly influence CT image noise characteristics and distribution.
• The findings provide insights into streak-like noise patterns and anisotropic noise correlations.
• Understanding these nonlinearities is essential for effective noise filtering and accurate tissue characterization in CT.