X-ray spectrum estimation for accurate attenuation simulation.

Area of Science:

• Medical Physics
• Spectroscopy
• Image Reconstruction

Background:

• Accurate X-ray spectra are crucial for quantitative analysis in various imaging applications.
• Existing spectrum estimation methods, like truncated singular value decomposition (TSVD) and expectation-maximization (EM), have limitations in accuracy and bias.

Purpose of the Study:

• To develop and validate a novel X-ray spectrum estimation method that enhances accuracy in simulating transmission through unknown attenuators.
• To minimize bias associated with prior spectral information.

Main Methods:

• Extended the truncated singular value decomposition (TSVD) algorithm by incorporating statistical transmission data properties and characteristic spectral peaks (k-edges).
• Introduced the prior truncated singular value decomposition (PTSVD) method, requiring minimal prior knowledge (energy positions of peaks/k-edges).
• Compared PTSVD against TSVD and expectation-maximization (EM) using simulated and measured transmission data.

Main Results:

• PTSVD demonstrated superior performance in spectral shape estimation and transmission data reproduction for noiseless simulated data compared to TSVD and EM.
• Even with increasing noise, PTSVD maintained its advantage in reproducing transmission data, though spectral shape estimation accuracy decreased.
• Measurement studies confirmed the findings from simulations, highlighting PTSVD's robustness.

Conclusions:

• The PTSVD method accurately estimates detected X-ray spectra, enabling reliable simulation of transmission through both known and unknown attenuators.
• PTSVD exhibits improved noise tolerance compared to TSVD and reduced bias compared to methods relying heavily on prior spectral information.
• This approach offers potential for smaller transmission errors, particularly with accurate transmission data.