Deep-learning-based denoising of X-ray differential phase and dark-field images

PURPOSE

Statistical photon noise has always been a common problem in X-ray multi-contrast imaging and significantly influenced the quality of retrieved differential phase and dark-field images. We intend to develop a deep learning-based denoising algorithm to reduce the noise of retrieved X-ray differential phase and dark-field images.

METHODS

A novel deep learning based image noise suppression algorithm (named DnCNN-P) is presented. We proposed two different denoising modes: Retrieval-Denoising mode (R-D mode) and Denoising-Retrieval mode (D-R mode). While the R-D mode denoises the retrieved images, the D-R mode denoises the raw phase stepping data. The two denoising modes are evaluated under different photon counts and visibilities.

RESULTS

Experimental results show that with the algorithm DnCNN-P used, the D-R mode always exhibits a better noise reduction under diverse experimental conditions, even in the case of a low photon count and/or a low visibility. With a detected photon count of 1800 and a visibility of 0.3, compared to the differential phase images without denoising, the standard deviation is reduced by 89.1% and 16.4% in the D-R and R-D modes. Compared to the dark-field images without denoising, the standard deviation is reduced by 83.7% and 12.6% in the D-R and R-D modes, respectively.

CONCLUSIONS

The novel supervised DnCNN-P algorithm can significantly reduce the noise in retrieved X-ray differential phase and dark-field images. We believe this novel algorithm can be a promising approach to improve the quality of X-ray differential phase and dark-field images, and therefore dose efficiency in future biomedical applications.

Self-aligned multi-layer X-ray absorption grating using large-area fabrication methods for X-ray phase-contrast imaging

X-ray phase-contrast (XPCi) imaging methods are an emerging medical imaging approach that provide significantly better soft tissue contrast and could function as a viable extension to conventional X-ray, CT, and even some MRI. Absorption gratings play a central role in grating-based XPCi systems, especially because they enable the acquisition of three images in a single exposure: transmission, refraction, and dark-field. An impediment to commercial development and adoption of XPCi imaging systems is the lack of large area, high aspect ratio absorption gratings. Grating technology development, primarily due to technological limitations, has lagged system development and today prevents the scaling up of XPCi system into a footprint and price point acceptable to the medical market. In this work, we report on a self-aligned multi-layer grating fabrication process that can enable large-area X-ray absorption gratings with micron-scale feature sizes. We leverage large-area fabrication techniques commonly employed by the thin-film transistor (TFT) display industry. Conventional ITO-on-glass substrates are used with a patterned film of Cr/Au/Cr that serves as a self-aligned lithography mask for backside exposure. Commonly available SU-8 photoresist is patterned using the backside exposure mask followed by an electroplating step to fill the gaps in the SU-8 with X-ray attenuating material. Consequently, the electroplated patterned material acts as a self-aligned photomask for subsequent SU-8 layer patterning and so forth. The repeatability of the reported process makes it suitable for achieving higher aspect ratio structures and is advantageous over previously reported X-ray LIGA approaches. A prototype three-layer grating, with a thickness of around [Formula: see text], having a visibility of 0.28 at [Formula: see text] with a [Formula: see text] active area was fabricated on a 4-inch glass substrate and demonstrated by modifying a commercially available 3D propagation-based XPCi Microscope. The scalable and cost-effective approach to build larger area X-ray gratings reported in this work can help expedite the commercial development and adoption of previously reported Talbot-Lau, speckle-tracking, as well as coded-aperture XPCi systems for large-area clinical and industrial applications.

Fabrication of X-ray absorption gratings by centrifugal deposition of bimodal tungsten particles in high aspect ratio silicon templates

Grating-based X-ray imaging employs high aspect ratio absorption gratings to generate contrast induced by attenuating, phase-shifting, and small-angle scattering properties of the imaged object. The fabrication of the absorption gratings remains a crucial challenge of the method on its pathway to clinical applications. We explore a simple and fast centrifugal tungsten particle deposition process into silicon-etched grating templates, which has decisive advantages over conventional methods. For that, we use a bimodal tungsten particle suspension which is introduced into a custom designed grating holder and centrifuged at over 1000×g. Gratings with 45 µm period, 450 µm depth, and 170 mm × 38 mm active area are successfully processed reaching a homogeneous absorber filling. The effective absorbing tungsten thickness in the trenches is 207 µm resulting in a filling ratio of 46.6% compared to a voidless filling. The grating was tested in a Talbot–Lau interferometer designed for clinical X-ray dark-field computed tomography, where visibilities up to 33.6% at 60 kV were achieved.

Retrieval of 3D information in X-ray dark-field imaging with a large field of view

X-ray dark-field imaging is a widely researched imaging technique, with many studies on samples of very different dimensions and at very different resolutions. However, retrieval of three-dimensional (3D) information for human thorax sized objects has not yet been demonstrated. We present a method, similar to classic tomography and tomosynthesis, to obtain 3D information in X-ray dark-field imaging. Here, the sample is moved through the divergent beam of a Talbot-Lau interferometer. Projections of features at different distances from the source seemingly move with different velocities over the detector, due to the cone beam geometry. The reconstruction of different focal planes exploits this effect. We imaged a chest phantom and were able to locate different features in the sample (e.g. the ribs, and two sample vials filled with water and air and placed in the phantom) to corresponding focal planes. Furthermore, we found that image quality and detectability of features is sufficient for image reconstruction with a dose of 68 μSv at an effective pixel size of [Formula: see text]. Therefore, we successfully demonstrated that the presented method is able to retrieve 3D information in X-ray dark-field imaging.

Whole-body x-ray dark-field radiography of a human cadaver

BACKGROUND

Grating-based x-ray dark-field and phase-contrast imaging allow extracting information about refraction and small-angle scatter, beyond conventional attenuation. A step towards clinical translation has recently been achieved, allowing further investigation on humans.

METHODS

After the ethics committee approval, we scanned the full body of a human cadaver in anterior-posterior orientation. Six measurements were stitched together to form the whole-body image. All radiographs were taken at a three-grating large-object x-ray dark-field scanner, each lasting about 40 s. Signal intensities of different anatomical regions were assessed. The magnitude of visibility reduction caused by beam hardening instead of small-angle scatter was analysed using different phantom materials. Maximal effective dose was 0.3 mSv for the abdomen.

RESULTS

Combined attenuation and dark-field radiography are technically possible throughout a whole human body. High signal levels were found in several bony structures, foreign materials, and the lung. Signal levels were 0.25 ± 0.13 (mean ± standard deviation) for the lungs, 0.08 ± 0.06 for the bones, 0.023 ± 0.019 for soft tissue, and 0.30 ± 0.02 for an antibiotic bead chain. We found that phantom materials, which do not produce small-angle scatter, can generate a strong visibility reduction signal.

CONCLUSION

We acquired a whole-body x-ray dark-field radiograph of a human body in few minutes with an effective dose in a clinical acceptable range. Our findings suggest that the observed visibility reduction in the bone and metal is dominated by beam hardening and that the true dark-field signal in the lung is therefore much higher than that of the bone.

Empirical beam hardening and ring artifact correction for x-ray grating interferometry (EBHC-GI)

PURPOSE

Talbot-Lau grating interferometry enables the use of polychromatic x-ray sources, extending the range of potential applications amenable to phase contrast imaging. However, these sources introduce beam hardening effects not only from the samples but also from the gratings. As a result, grating inhomogeneities due to manufacturing imperfections can cause spectral nonuniformity artifacts when used with polychromatic sources. Consequently, the different energy dependencies of absorption, phase, and visibility contrasts impose challenges that so far have limited the achievable image quality. The purpose of this work was to develop and validate a correction strategy for grating-based x-ray imaging that accounts for beam hardening generated from both the imaged object and the gratings.

METHODS

The proposed two-variable polynomial expansion strategy was inspired by work performed to address beam hardening from a primary modulator. To account for the multicontrast nature of grating interferometry, this approach was extended to each contrast to obtain three sets of correction coefficients, which were determined empirically from a calibration scan. The method's feasibility was demonstrated using a tabletop Talbot-Lau grating interferometer micro-computed tomography (CT) system using CT acquisitions of a water sample and a silicon sample, representing low and high atomic number materials. Spectral artifacts such as cupping and ring artifacts were quantified using mean squared error (MSE) from the beam-hardening-free target image and standard deviation within a reconstructed image of the sample. Finally, the model developed using the water sample was applied to a fixated murine lung sample to demonstrate robustness for similar materials.

RESULTS

The water sample's absorption CT image was most impacted by spectral artifacts, but following correction to decrease ring artifacts, an 80% reduction in MSE and 57% reduction in standard deviation was observed. The silicon sample created severe artifacts in all contrasts, but following correction, MSE was reduced by 94% in absorption, 96% in phase, and 90% in visibility images. These improvements were due to the removal of ring artifacts for all contrasts and reduced cupping in absorption and phase images and reduced capping in visibility images. When the water calibration coefficients were applied to the lung sample, ring artifacts most prominent in the absorption contrast were eliminated.

CONCLUSIONS

The described method, which was developed to remove artifacts in absorption, phase, and normalized visibility micro-CT images due to beam hardening in the system gratings and imaged object, reduced the MSE by up to 96%. The method depends on calibrations that can be performed on any system and does not require detailed knowledge of the x-ray spectrum, detector energy response, grating attenuation properties and imperfections, or the geometry and composition of the imaged object.

Imaging characteristics of intravascular spherical contrast agents for grating-based x-ray dark-field imaging – effects of concentrations, spherical sizes and applied voltage

This study investigates the x-ray scattering characteristics of microsphere particles in x-ray-grating-based interferometric imaging at different concentrations, bubble sizes and tube voltages (kV). Attenuation (ATI), dark-field (DFI) and phase-contrast (PCI) images were acquired. Signal-to-noise (SNR) and contrast-to-noise ratios with water (CNRw) and air as reference (CNRa) were determined. In all modalities, a linear relationship between SNR and microbubbles concentration, respectively, microsphere size was found. A significant gain of SNR was found when varying kV. SNR was significantly higher in DFI and PCI than ATI. The highest gain of SNR was shown at 60 kV for all media in ATI and DFI, at 80 kV for PCI. SNR for all media was significantly higher compared to air and was slightly lower compared to water. A linear relationship was found between CNRa, CNRw, concentration and size. With increasing concentration and decreasing size, CNRa and CNRw increased in DFI, but decreased in PCI. Best CNRa and CNRw was found at specific combination of kV and concentration/size. Highest average CNRa and CNRw was found for microspheres in ATI and PCI, for microbubbles in DFI. Microspheres are a promising contrast-media for grating-based-interferometry, if kV, microsphere size and concentration are appropriately combined.

Wave optics simulation of grating-based X-ray phase-contrast imaging using 4D Mouse Whole Body (MOBY) phantom

PURPOSE

Demonstrate realistic simulation of grating-based x-ray phase-contrast imaging (GB-XPCI) using wave optics and the four-dimensional Mouse Whole Body (MOBY) phantom defined with non-uniform rational B-splines (NURBS).

METHODS

We use a full-wave approach, which uses wave optics for x-ray wave propagation from the source to the detector. This forward imaging model can be directly applied to NURBS-defined numerical phantoms such as MOBY. We assign the material properties (attenuation coefficient and electron density) of each model part using the data for adult human tissues. The Poisson noise is added to the simulated images based on the calculated photon flux at each pixel.

RESULTS

We simulate the intensity images of the MOBY phantom for eight different grating positions. From the simulated images, we calculate the absorption, differential phase, and normalized visibility contrast images. We also predict how the image quality is affected by different exposure times.

CONCLUSIONS

GB-XPCI can be simulated with the full-wave approach and a realistic numerical phantom defined with NURBS.

Proof-of-concept Talbot-Lau x-ray interferometry with a high-intensity, high-repetition-rate, laser-driven K-alpha source

Talbot-Lau x-ray interferometry is a grating-based phase-contrast technique, which enables measurement of refractive index changes in matter with micrometric spatial resolution. The technique has been established using a variety of hard x-ray sources, including synchrotron, free-electron lasers, and x-ray tubes, and could be used in the optical range for low-density plasmas. The tremendous development of table-top high-power lasers makes the use of high-intensity, laser-driven K-alpha sources appealing for Talbot-Lau interferometer applications in both high-energy-density plasma experiments and biological imaging. To this end, we present the first, to the best of our knowledge, feasibility study of Talbot-Lau phase-contrast imaging using a high-repetition-rate laser of moderate energy (100 mJ at a repetition rate of 10 Hz) to irradiate a copper backlighter foil. The results from up to 900 laser pulses were integrated to form interferometric images. A constant fringe contrast of 20% is demonstrated over 100 accumulations, while the signal-to-noise ratio continued to increase with the number of shots. Phase retrieval is demonstrated without prior ex-situ phase stepping. Instead, correlation matrices are used to compensate for the displacement between reference acquisition and the probing of a PMMA target rod. The steps for improved measurements with more energetic laser systems are discussed. The final results are in good agreement with the theoretically predicted outcomes, demonstrating the applicability of this diagnostic to a range of laser facilities for use across several disciplines.

Modeling the Moiré fringe visibility of Talbot-Lau X-ray grating interferometry for single-frame multi-contrast imaging

Moiré fringe method in X-ray grating interferometry is characterized by its advantage to obtain multi-contrast data through single-frame imaging. However, how the visibility of the Moiré fringe is influenced by the system parameters, such as the misalignment angle, still lacks investigation, although it closely relates to the signal-to-noise ratio of the image data. In this paper, a simplified model of the Moiré fringe visibility is presented, which reveals that the Moiré fringe visibility can be expressed as the product of a misalignment-angle-related "sinc" function and a relatively independent factor. The following experimental results further suggest that the crosstalk between the detector pixels in the direction perpendicular to the Moiré fringe orientation is another main cause for the visibility reduction of the Moiré fringes.

Recent Progress in X-ray and Neutron Phase Imaging with Gratings

Under the JST-ERATO project in progress to develop X-ray and neutron phase-imaging methods together, recent achievements have been selected and reviewed after describing the merit and the principle of the phase imaging method. For X-ray phase imaging, recent developments of four-dimensional phase tomography and phase microscopy at SPring-8, Japan are mainly presented. For neutron phase imaging, an approach in combination with the time-of-flight method developed at J-PARC, Japan is described with the description of new Gd grating fabrication.

The trigonometric orthogonality of phase-stepping curves in grating-based x-ray phase-contrast imaging: Integral property and its implications for noise optimization

PURPOSE

Grating-based x-ray phase-contrast imaging (GPCI) is a promising technique for clinical applications as it can provide two newly emerging imaging modalities (differential phase-contrast and dark-field contrast) in addition to the conventional absorption contrast. As far, phase-stepping strategy is the most commonly used approach in GPCI to indirectly acquire differential phase-contrast and dark-field contrast. It is known that the obtained phase-stepping curves (PSCs) have the cosine property and the convolution property, leading to two types of information retrieval approaches in literature: the Fourier component analysis and the multi-order moment analysis. The purpose of this paper is to derive a new property of PSCs and apply the property to noise optimization for information retrieval.

METHODS

Based on the cosine expression of the flat PSC without the sample and the well-established convolution relationship between the flat PSC and the sample PSC, we reveal an important integral property of PSCs: the inner product of PSCs and an arbitrary function contains only zero-order and first-order components in the Fourier series. Furthermore, we apply the property to the direct multi-order moment analysis and propose a set of generalized forms including an optimal one in the presence of noise.

RESULTS

To validate the effectiveness of our analysis, we compared the simulated and real experiment results retrieved by the original direct multi-order moment analysis with the ones retrieved by our proposed noise-optimal form. A significant improvement of noise performance by our method is observed and the improvement ratio in differential phase-contrast is consistent with our theoretical calculation (39.2%).

CONCLUSIONS

In this paper, we reveal a new integral property of the acquired PSCs with and without samples in GPCI, which can be applied to information retrieval approaches like the direct multi-order moment analysis. Then we optimize these approaches to improve the noise performance, offering great potentials of dose reduction in practical applications.

Contrast-to-noise ratios and thickness-normalized, ventilation-dependent signal levels in dark-field and conventional in vivo thorax radiographs of two pigs

Lung tissue causes significant small-angle X-ray scattering, which can be visualized with grating-based X-ray dark-field imaging. Structural lung diseases alter alveolar microstructure, which often causes a dark-field signal decrease. The imaging method provides benefits for diagnosis of such diseases in small-animal models, and was successfully used on porcine and human lungs in a fringe-scanning setup. Micro- and macroscopic changes occur in the lung during breathing, but their individual effects on the dark-field signal are unknown. However, this information is important for quantitative medical evaluation of dark-field thorax radiographs. To estimate the effect of these changes on the dark-field signal during a clinical examination, we acquired in vivo dark-field chest radiographs of two pigs at three ventilation pressures. Pigs were used due to the high degree of similarity between porcine and human lungs. To analyze lung expansion separately, we acquired CT scans of both pigs at comparable posture and ventilation pressures. Segmentation, masking, and forward-projection of the CT datasets yielded maps of lung thickness and logarithmic lung attenuation signal in registration with the dark-field radiographs. Upon correlating this data, we discovered approximately linear relationships between the logarithmic dark-field signal and both projected quantities for all scans. Increasing ventilation pressure strongly decreased dark-field extinction coefficients, whereas the ratio of lung dark-field and attenuation signal changed only slightly. Furthermore, we investigated ratios of dark-field and attenuation noise levels at realistic signal levels via calculations and phantom measurements. Dark-field contrast-to-noise ratio (CNR) per lung height was 5 to 10% of the same quantity in attenuation. We conclude that better CNR performance in the dark-field modality is typically due to greater anatomical noise in the conventional radiograph. Given the high physiological similarity of human and porcine lungs, the presented thickness-normalized, ventilation-dependent values allow estimation of dark-field activity of human lungs of variable size and inspiration, which facilitates the design of suitable clinical imaging setups.

Exploration of different x-ray Talbot-Lau setups for dark-field lung imaging examined in a porcine lung

X-ray dark-field imaging is a promising technique for lung diagnosis. Due to the alveolar structure of lung tissue, a higher contrast is obtained by the dark-field image compared to the attenuation image. Animal studies indicate an enhancement regarding the detection of lung diseases in early stages. In this publication, we focus on the influence of different Talbot-Lau interferometer specifications while maintaining the x-ray source, sample magnification and detector system. By imaging the same porcine lung with three different grating sets, we analyze the contrast-to-noise ratio of the obtained dark-field images with respect to visibility and correlation length. We demonstrate that relatively large grating periods of the phase and of the analyzer grating are sufficient for high quality lung imaging at reasonable dose levels.

Single-shot Talbot-Lau x-ray dark-field imaging of a porcine lung applying the moiré imaging approach

Talbot-Lau x-ray imaging provides additionally to the conventional attenuation image, two further images: the differential phase-contrast image which is especially sensitive to differences in refractive properties and the dark-field image which is showing the x-ray scattering properties of the object. Thus, in the dark-field image sub-pixeled object information can be observed. As it has been shown in recent studies, this is of special interest for lung imaging. Changes in the alveoli structure, which are in the size of one detector pixel, can be seen in the dark-field images. A fast acquisition process is crucial to avoid motion artifacts due to heartbeat and breathing of the patient. Using moiré imaging the images can be acquired with a single-shot exposure. Nevertheless, the spatial resolution is reduced compared to the phase-stepping acquisition. We evaluate the results of both imaging techniques towards their feasibility in clinical routine. Furthermore, we analyse the influence of artificial linear object movement on the image quality, in order to simulate the heartbeat of a patient.