Enhancing the X-Ray Differential Phase Contrast Image Quality With Deep Learning Technique

Objective: The purpose of this work is to investigate the feasibility of using deep convolutional neural network (CNN) to improve the image quality of a grating-based X-ray differential phase contrast imaging (XPCI) system. Methods: In this work, a novel deep CNN based phase signal extraction and image noise suppression algorithm (named as XP-NET) is developed. The numerical phase phantom, the ex vivo biological specimen and the ACR breast phantom are evaluated via the numerical simulations and experimental studies, separately. Moreover, images are also evaluated under different low radiation levels to verify its dose reduction capability. Results: Compared with the conventional analytical method, the novel XP-NET algorithm is able to reduce the bias of large DPC signals and hence increasing the DPC signal accuracy by more than 15%. Additionally, the XP-NET is able to reduce DPC image noise by about 50% for low dose DPC imaging tasks. Conclusion: This proposed novel end-to-end supervised XP-NET has a great potential to improve the DPC signal accuracy, reduce image noise, and preserve object details. Significance: We demonstrate that the deep CNN technique provides a promising approach to improve the grating-based XPCI performance and its dose efficiency in future biomedical applications.

I. Introduction

As a novel X-ray imaging method, the Talbot-Lau grating interferometry enabled X-ray phase contrast imaging (XPCI) method [1] can simultaneously capture three unique object information from one set of acquired data: absorption, differential phase contrast (DPC), and dark-field (DF). Compared with the conventional absorption contrast, studies have found that the DPC signal could provide superior contrast performance for certain types of soft tissues and other low-density, low-Z materials [2], [3]. Besides, the DF signal is found to be particularly sensitive to certain fine structures such as micro-calcification inside breast tissue [4]–[6]. Due to these potential advantages, numerous research interests [7]–[12] have been attracted with the hope to translate such novel X-ray imaging method into biomedical and clinical applications. Despite of these potential advancements, however, the low radiation dose efficiency of the Talbot-Lau interferometer strongly impedes its wide applications. This is mainly due to the photon absorption on the analyzer grating. Usually, the analyzer grating blocks more than half of the photons that have already penetrated through the object. To overcome such difficulty, the imaging performance of a grating based XPCI system needs to be improved.