arXiv subjects

Simon Zabler

Publications and source records attributed to Simon Zabler.

7 recordsLinked to original sources

arXiv2025

Multispectral CT Denoising via Simulation-Trained Deep Learning: Experimental Results at the ESRF BM18

Multispectral computed tomography (CT) enables advanced material characterization by acquiring energy-resolved projection data. However, since the incoming X-ray flux is be distributed across multiple narrow energy bins, the photon count per bin is greatly reduced compared to standard energy-integrated imaging. This inevitably introduces substantial noise, which can either prolong acquisition times and make scan durations infeasible or degrade image quality with strong noise artifacts. To address this challenge, we present a dedicated neural network-based denoising approach tailored for multispectral CT projections acquired at the BM18 beamline of the ESRF. The method exploits redundancies across angular, spatial, and spectral domains through specialized sub-networks combined via stacked generalization and an attention mechanism. Non-local similarities in the angular-spatial domain are leveraged alongside correlations between adjacent energy bands in the spectral domain, enabling robust noise suppression while preserving fine structural details. Training was performed exclusively on simulated data replicating the physical and noise characteristics of the BM18 setup, with validation conducted on CT scans of custom-designed phantoms containing both high-Z and low-Z materials. The denoised projections and reconstructions demonstrate substantial improvements in image quality compared to classical denoising methods and baseline CNN models. Quantitative evaluations confirm that the proposed method achieves superior performance across a broad spectral range, generalizing effectively to real-world experimental data while significantly reducing noise without compromising structural fidelity.

eess.IV↗

arXiv2020

Review and experimental verification of X-ray darkfield signal interpretations with respect to quantitative isotropic and anisotropic darkfield computed tomography

Talbot(-Lau) interferometric X-ray darkfield imaging has, over the past decade, gained substantial interest for its ability to provide insights into a sample's microstructure below the imaging resolution by means of ultra small angle scattering effects. Quantitative interpretations of such images depend on models of the signal origination process that relate the observable image contrast to underlying physical processes. A review of such models is given here and their relation to the wave optical derivations by Yashiro et al. and Lynch et al. as well as to small angle X-ray scattering is discussed. Fresnel scaling is introduced to explain the characteristic distance dependence observed in cone beam geometries. Moreover, a model describing the anisotropic signals of fibrous objects is derived. The Yashiro-Lynch model is experimentally verified both in radiographic and tomographic imaging in a monochromatic synchrotron setting, considering both the effects of material and positional dependence of the resulting darkfield contrast. The effect of varying sample-detector distance on the darkfield signal is shown to be non-negligible for tomographic imaging, yet can be largely compensated for by symmetric acquisition trajectories. The derived orientation dependence of the darkfield contrast of fibrous materials both with respect to variations in autocorrelation width and scattering cross section is experimentally validated using carbon fiber reinforced rods.

physics.app-ph↗

arXiv2020

Local Fiber Orientation from X-ray Region-of-Interest Computed Tomography of large Fiber Reinforced Composite Components

The local fiber orientation is a micro-structural feature crucial for the mechanical properties of parts made from fiber reinforced polymers. It can be determined from micro-computed tomography data and subsequent quantitative analysis of the resulting 3D images. However, although being by nature non-destructive, this method so far has required to cut samples of a few millimeter edge length in order to achieve the high lateral resolution needed for the analysis. Here, we report on the successful combination of region-of-interest scanning with structure texture orientation analysis rendering the above described approach truly non-destructive. Several regions of interest in a large bearing part from the automotive industry made of fiber reinforced polymer are scanned and analyzed. Differences of these regions with respect to local fiber orientation are quantified. Moreover, consistency of the analysis based on scans at varying lateral resolutions is proved. Finally, measured and numerically simulated orientation tensors are compared for one of the regions.

eess.IV↗

arXiv2019

Comparing Image Quality in Phase Contrast sub$μ$ X-Ray Tomography -- A Round-Robin Study

How to evaluate and compare image quality from different sub-micrometer (sub$μ$) CT scans? A simple test phantom made of polymer microbeads is used for recording projection images as well as 13 CT scans in a number of commercial and non-commercial scanners. From the resulting CT images, signal and noise power spectra are modeled for estimating volume signal-to-noise ratios (3D SNR spectra). Using the same CT images, a time- and shape-independent transfer function (MTF) is computed for each scan, including phase contrast effects and image blur ($\mathrm{MTF_{blur}}$). The SNR spectra and MTF of the CT scans are compared to 2D SNR spectra of the projection images. In contrary to 2D SNR, volume SNR can be normalized with respect to the object's power spectrum, yielding detection effectiveness (DE) a new measure which reveals how technical differences as well as operator-choices strongly influence scan quality for a given measurement time. Using DE, both source-based and detector-based sub$μ$ CT scanners can be studied and their scan quality can be compared. Future application of this work requires a particular scan acquisition scheme which will allow for measuring 3D signal-to-noise ratios, making the model fit for 3D noise power spectra obsolete.

physics.ins-det↗

arXiv2019

SNR Spectra as a Quantitative Model for Image Quality in Polychromatic X-Ray Imaging

In polychromatic x-ray imaging for nondestructive testing, material science or medical applications, image quality is usually a problem of detecting sample structure in noisy data. This problem is typically stated this way: As many photons as possible need to be detected to get a good image quality. We instead propose to use the concept of signal detection, which is more universal. In signal detection, it is the sample properties which are detected. Photons play the role of information carriers for the signal. Signal detection for example allows modeling the effects which polychromaticity has on image quality. $\mathit{SNR}$ spectra (= spatial $\mathit{SNR}$) are used as a quantity to describe if reliable signal detection is possible. They include modulation transfer and phase contrast in addition to noisiness effects. $\mathit{SNR}$ spectra can also be directly measured, which means that theoretical predictions can easily be tested. We investigate the effects of signal and noise superposition on the $\mathit{SNR}$ spectrum and show how selectively not detecting photons can increase the image quality.

physics.ins-det↗

arXiv2018

Optimization based evaluation of grating interferometric phase stepping series and analysis of mechanical setup instabilities

The diffraction contrast modalities accessible by X-ray grating interferometers are not imaged directly but have to be inferred from sine like signal variations occurring in a series of images acquired at varying relative positions of the interferometer's gratings. The absolute spatial translations involved in the acquisition of these phase stepping series usually lie in the range of only a few hundred nanometers, wherefore positioning errors as small as 10nm will already translate into signal uncertainties of one to ten percent in the final images if not accounted for. Classically, the relative grating positions in the phase stepping series are considered input parameters to the analysis and are, for the Fast Fourier Transform that is typically employed, required to be equidistantly distributed over multiples of the gratings' period. More general optimization based approaches with relaxed requirements on the sampling positions are often deemed too time consuming in contrast. In the following, a fast converging optimization scheme is presented simultaneously determining the phase stepping curves' parameters as well as the actually performed motions of the stepped grating, including also erroneous rotational motions which are commonly neglected. While the correction of solely the translational errors along the stepping direction is found to be sufficient with regard to the reduction of image artifacts, the possibility to also detect minute rotations about all axes proves to be a valuable tool for system calibration and monitoring. The simplicity of the provided algorithm, in particular when only considering translational errors, makes it well suitable as a standard evaluation procedure also for large image series.

physics.ins-det↗

arXiv2012

Comparison of different sources for laboratory X-ray microscopy

This paper describes the setup of two different solutions for laboratory X-ray microscopy working with geometric magnification. One setup uses thin-film transmission targets with an optimized tungsten-layer thickness and the electron gun and optics of an electron probe micro analyzer to generate a very small X-ray source. The other setup is based on a scanning electron microscope and uses microstructured reflection targets. We also describe the structuring process for these targets. In both cases we show that resolutions of 100 nm can be achieved. Also the possibilities of computed tomography for 3D imaging are explored and we show first imaging examples of high-absorption as well as low-absorption specimens to demonstrate the capabilities of the setups.

physics.ins-det↗