Increasing the darkfield contrast-to-noise ratio using a deconvolution-based information retrieval algorithm in X-ray...

Area of Science:

• Medical Imaging
• Physics

Background:

• X-ray grating-based phase-contrast imaging offers enhanced sensitivity over conventional radiography.
• Existing methods, like Fourier analysis, are widely used but may have limitations in noise reduction and dose efficiency.
• Phase stepping curve (PSC) analysis provides an alternative approach for image reconstruction.

Purpose of the Study:

• To investigate a novel deconvolution algorithm for X-ray grating-based phase-contrast imaging.
• To compare the performance of the deconvolution method with the conventional Fourier analysis method.
• To evaluate the contrast-to-noise ratio (CNR) and dose-saving potential of the novel algorithm.

Main Methods:

• Application of a deconvolution algorithm based on object and reference PSCs (Modregger et al.) to polychromatic X-ray data.
• Comparison with results obtained using a standard Fourier analysis method.
• Experimental estimation of dose saving potential and CNR improvement.

Main Results:

• Both deconvolution and Fourier methods yielded identical results for absorption and differential phase images.
• The deconvolution method demonstrated a 1.17-fold increase in mean contrast-to-noise ratio (CNR) for darkfield images.
• The deconvolution method requires 1.66 times less dose than the conventional method to achieve similar CNR.

Conclusions:

• The deconvolution method offers superior background noise properties, leading to improved CNR and dose efficiency in X-ray phase-contrast imaging.
• The enhanced performance is particularly evident in darkfield imaging.
• A trade-off exists, as the deconvolution method's results can be dependent on photon statistics, affecting inter-measurement comparability at different dose levels.