A quantitative correction for phase wrapping artifacts in hard X-ray grating interferometry

X-ray grating interferometry-based computed tomography is a phase contrast imaging technique that provides non-destructive, quantitative, and three-dimensional visualization with contrast superior to traditional absorption-based techniques, especially for materials primarily composed of low Z elements, such as biological tissues. However, it relies on measurements of the lateral shift of an interference pattern and is thus susceptible to so-called phase wrapping artifacts, which mainly occur at the sample-air interface. In this work, we present an algorithm for removal of such artifacts in the case of cylindrical samples and an experiment to verify its accuracy. The proposed algorithm is applied to the sinogram after phase retrieval and prior to reconstruction by finding sample edges with the absorption sinogram and replacing regions of the phase wrapped sinogram with modeled data. Our measurements show that the algorithm removes artifacts and produces more accurate δ values, as validated by measurements without phase wrapping. Our correction algorithm allows for measurements without submerging the sample in a water bath, simplifying the experimental setup and avoiding motion artifacts from gas bubbles.X-ray grating interferometry-based computed tomography is a phase contrast imaging technique that provides non-destructive, quantitative, and three-dimensional visualization with contrast superior to traditional absorption-based techniques, especially for materials primarily composed of low Z elements, such as biological tissues. However, it relies on measurements of the lateral shift of an interference pattern and is thus susceptible to so-called phase wrapping artifacts, which mainly occur at the sample-air interface. In this work, we present an algorithm for removal of such artifacts in the case of cylindrical samples and an experiment to verify its accuracy. The proposed algorithm is applied to the sinogram after phase retrieval and prior to reconstruction by finding sample edges with the absorption sinogram and replacing regions of the phase wrapped sinogram with modeled data. Our measurements show that the algorithm removes artifacts and produces more accurate δ values, as validated by measurements ...