Phase retrieval in in-line x-ray phase contrast imaging based on total variation minimization

Area of Science:

• Physics
• Medical Imaging
• Computational Science

Background:

• Current X-ray phase-contrast imaging phase retrieval methods solve underdetermined linear systems.
• Tikhonov regularization (L2-norm) is commonly used but has limitations.
• Need for improved phase retrieval techniques in X-ray imaging.

Purpose of the Study:

• To present a novel phase retrieval approach using Total Variation (TV) minimization.
• To compare TV minimization with Tikhonov regularization for deconvolution in phase retrieval.
• To evaluate performance on simulated and experimental X-ray phase-contrast imaging data.

Main Methods:

• Incorporation of Total Variation (TV) minimization for deconvolution.
• Application to common linear phase-contrast models in X-ray imaging.
• Comparative analysis against Tikhonov regularized deconvolution.

Main Results:

• TV minimization demonstrated improved accuracy in phase retrieval reconstructions.
• Enhanced performance observed for both single and multiple distance phase-contrast images.
• The method showed robustness against noise and nonlinear imaging effects.

Conclusions:

• Total Variation (TV) minimization offers a superior alternative to Tikhonov regularization for X-ray phase retrieval.
• The proposed method enhances image reconstruction quality in challenging imaging conditions.
• This technique advances propagation-based X-ray phase-contrast imaging capabilities.