Effectiveness of X-ray phase-contrast tomography: effects of pixel size and magnification on image noise

Abstract

Noise magnitude in conventional attenuation X-ray tomography (CT) is strongly dependent on the pixel size and/or the geometrical magnification, thereby limiting the possibility of achieving high-resolution low-dose CT imaging. In this context, the use of Propagation-Based Imaging (PBI) phase-contrast technique coupled with the application of a suitable Phase-Retrieval (PhR) filter is a valuable tool to overcome such limitation. In fact, at fixed radiation dose, the noise dependence on the effective pixel size when the PhR filter is applied is much shallower with respect to conventional CT imaging. Making use of a theoretical framework developed by other authors, this work demonstrates quantitatively the dependence of CT image noise on pixel size and magnification in PBI. Calculations are compared with experimental images of a breast specimen imaged at the SYRMEP beamline at the Elettra synchrotron facility (Trieste, Italy), with a CdTe photon-counting detector in PBI configuration. The results, expressed in terms of Signal-to-Noise Ratio (SNR) gain due to the PhR application, show a good agreement between predictions and experimental data at all pixel pitches and magnifications, quantitatively demonstrating the importance of going towards detectors featuring smaller pixels (or higher spatial resolution) to fully exploit the advantages of PBI and PhR. Specifically, SNR gain up to a factor of 20 is observed at the smallest pixel pitch (60 μ m) and largest magnification (1.40). At the same time, as predicted theoretically, larger magnifications correspond to lower image noise (or higher SNR) when PhR is applied: this trend is unparalleled in attenuation-based CT imaging where larger magnifications, hence smaller effective pixel sizes, lead to a higher noise.