Revised noise model of dark-field imaging using a grating interferometer

Abstract

Inspired by recent experimental results, we report on a revised noise model of dark-field imaging using a grating interferometer. The accurate knowledge of the noise of the dark-field image is essential for noise reduction and image quality improvement. A revised noise model is theoretically derived in terms of the standard deviation of the dark-field image. While a low visibility assumption was made in previous models, the revised noise model makes no assumption on the visibility. It is​ shown by synchrotron radiation experimental results that, with a high visibility, the revised noise model agreed quantitatively with experimentally measured values, while previous models overestimated the noise present in the dark-field image significantly. Therefore, the revised model provides an accurate description of the noise behavior of grating-based dark-field imaging, and can be of use for energy-resolved dark-field imaging and potential dose optimizations.