Grating-based x-ray differential phase contrast imaging with twin peaks in phase-stepping curves-phase retrieval and dewrapping

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.

Direct information retrieval after 3D reconstruction in grating-based X-ray phase-contrast computed tomography

Grating-based X-ray differential phase-contrast imaging has attracted a great amount of attention and has been considered as a potential imaging method in clinical medicine because of its compatibility with the traditional X-ray tube source and the possibility of a large field of view. Moreover, phase-contrast computed tomography provides three-dimensional phase-contrast visualization. Generally, two-dimensional information retrieval performed on every projection is required prior to three-dimensional reconstruction in phase-contrast computed tomography. In this paper, a three-dimensional information retrieval method to separate absorption and phase information directly from two reconstructed images is derived. Theoretical derivations together with numerical simulations have been performed to confirm the feasibility and veracity of the proposed method. The advantages and limitations compared with the reverse projection method are also discussed. Owing to the reduced data size and the absence of a logarithm operation, the computational time for information retrieval is shortened by the proposed method. In addition, the hybrid three-dimensional images of absorption and phase information were reconstructed using an absorption reconstruction algorithm, hence the existing data pre-processing methods and iterative reconstruction algorithms in absorption reconstruction may be utilized in phase reconstruction immediately.

Improved reconstruction of phase-stepping data for Talbot-Lau x-ray imaging

Grating-based Talbot-Lau x-ray interferometry is a popular method for measuring absorption, phase shift, and small-angle scattering. The standard acquisition method for this modality is phase stepping, where the Talbot pattern is reconstructed from multiple images acquired at different grating positions. We review the implicit assumptions in phase-stepping reconstruction, and find that the assumptions of perfectly known grating positions and homoscedastic noise variance are violated in some scenarios. Additionally, we investigate a recently reported estimation bias in the visibility and dark-field signal. To adapt the phase-stepping reconstruction to these findings, we propose three improvements to the reconstruction. These improvements are (a) to use prior knowledge to compute more accurate grating positions to reduce moiré artifacts, (b) to utilize noise variance information to reduce dark-field and phase noise in high-visibility acquisitions, and (c) to perform correction of an estimation bias in the interferometer visibility, leading to more quantitative dark-field imaging in acquisitions with a low signal-to-noise ratio. We demonstrate the benefit of our methods on simulated data, as well as on images acquired with a Talbot-Lau interferometer.

Reducing radiation dose in grating based x-ray phase contrast CT with twin-peaks in its phase stepping curves

PURPOSE

Differential phase contrast CT has been recognized as an x-ray imaging method with the potential to greatly improve the differentiation of soft tissues. Talbot interferometry has been one of the promising solutions allowing implementation with commercially available x-ray tubes with a polychromatic spectrum. Mainly due to imperfections in grating fabrication and the polychromatic spectrum of x-ray beam, a twin-peaks phenomenon may exist in phase stepping curves (PSCs) and degrade the performance of phase retrieval. The authors have previously proposed a Fourier analysis based method for phase retrieval in the scenario wherein the twin-peaks phenomenon occurs in PSCs. In this work, the authors propose a 5-step algebraic method for phase retrieval and investigate the potential of reducing radiation dose while both the Fourier and algebraic methods are being utilized for phase retrieval.

METHODS

The algebraic method to deal with the twin-peaks phenomenon, in which a set of linear equations with five unknown variables is needed for phase retrieval, is an extension of the so-called 3-step method that has been used in the scenario wherein only single-peak exists in the PSCs. In addition to a numerical phantom, two sets of experimental data (a phantom made of organic materials and a small animal) acquired by a prototype differential phase contrast CT system are employed to evaluate the performance of the Fourier and algebraic phase retrieval methods and their potential in radiation dose reduction.

RESULTS

The evaluation by both numerical phantom and experimental data shows that the algebraic method works as well as the Fourier method in phase retrieval if the twin-peaks phenomenon in the PSCs is appropriately dealt with. In addition, while the radiation dose associated with data acquisition is being reduced via fewer phase shifting steps, the algebraic method can maintain a better performance compared to the Fourier method.

CONCLUSIONS

Along with the Fourier method, the proposed 5-step algebraic method can cope with the twin-peaks phenomenon in phase retrieval. With decreased phase shifting steps and thus radiation dose, the proposed algebraic method performs better than the Fourier method, providing a practical solution for implementation of grating based differential phase contrast CT.