arXiv subjects

Fabian Wagner

Publications and source records attributed to Fabian Wagner.

At least 19 recordsLinked to original sources

arXiv2026

Charting causal set configuration space with graph observables

The configuration space of causal sets is vast. It is a critical goal to map out this space. Here, we take a practical step towards this goal. We investigate nine classes of causal sets, most of them not studied before. These include manifoldlike causal sets with inhomogeneous Ricci curvature, both topologically trivial and nontrivial. We also study classes of non-manifoldlike causal sets, including lattices, layered orders as well as Lorentzian quasicrystals. Finally, we study classes of causal sets that are not manifoldlike, but are expected to become manifoldlike under a suitable coarse-graining process. We use this broad range of distinct classes of causal sets as a testbed for observables. Rather than focusing on continuum-geometry inspired observables, such as curvature invariants, which often exhibit large fluctuations and are computationally very expensive, we focus on graph observables, including some observables that constitute subgraph statistics and some that are global. We find that three observables, namely the link degree distribution, the eigenvalues of the graph Laplacian of the symmetrized Hasse diagram and the recently proposed abundance of causal intervals, can distinguish between the distinct classes of causal sets. This is made possible by the small fluctuations that these observables have in most classes.

gr-qc↗

arXiv2026

Filter2Noise: A Framework for Interpretable and Zero-Shot Low-Dose CT Image Denoising

Noise in low-dose computed tomography (LDCT) can obscure important diagnostic details. While deep learning offers powerful denoising, supervised methods require impractical paired data, and self-supervised alternatives often use opaque, parameter-heavy networks that limit clinical trust. We propose Filter2Noise (F2N), a novel self-supervised framework for interpretable, zero-shot denoising from a single LDCT image. Instead of a black-box network, its core is an Attention-Guided Bilateral Filter, a transparent, content-aware mathematical operator. A lightweight attention module predicts spatially varying filter parameters, making the process transparent and allowing interactive radiologist control. To learn from a single image with correlated noise, we introduce a multi-scale self-supervised loss coupled with Euclidean Local Shuffle (ELS) to disrupt noise patterns while preserving anatomical integrity. On the Mayo Clinic LDCT Challenge, F2N achieves state-of-the-art results, outperforming competing zero-shot methods by up to 3.68 dB in PSNR. It accomplishes this with only 3.6k parameters, orders of magnitude fewer than competing models, which accelerates inference and simplifies deployment. By combining high performance with transparency, user control, and high parameter efficiency, F2N offers a trustworthy solution for LDCT enhancement. We further demonstrate its applicability by validating it on clinical photon-counting CT data. Code is available at: https://github.com/sypsyp97/Filter2Noise.

eess.IV↗

arXiv2025

Universality of free fall in Planck-scale deformed Newtonian gravity

The universality of free fall is one of the most cherished principles in classical gravity. Its fate in the quantum world is one of the key questions in fundamental physics. We investigate the universality of free fall in the context of Planck scale modifications of Newtonian gravity. Starting from a doubly-special-relativity setting we take the Newtonian limit to obtain deformed Galilean relativity. We study the interaction between two test particles, subject to deformed Galilean relativity, and a classical, undeformed gravitational source, the Earth. Such an interaction is investigated here for the first time. Considering the two test particles falling freely in the source's gravitational field, we examine whether the universality of free fall is affected by deformed relativistic symmetries. We show that, in general, the universality of free fall is violated. Remarkably, we find that there exist distinguished models for which the universality of free fall is realized and which predict a specific modification of the Newtonian potential.

gr-qc↗

arXiv2025

Regulator and gauge dependence of the Abelian gauge coupling in asymptotically safe quantum gravity

Both General Relativity and the Standard Model of particle physics are not UV complete. General Relativity is perturbatively non-renormalizable, while the Standard Model features Landau poles, where couplings are predicted to diverge at finite energies, e.g., in the Abelian gauge sector. Asymptotically safe quantum gravity may resolve both of these issues at the same time. In this paper, we assess the systematic uncertainties associated with this scenario, in particular with the gravitationally induced UV-completion of the Abelian gauge sector. Specifically, we study the dependence of this qualitative feature, namely the existence of a UV-complete gauge sector, on unphysical choices like the gauge, and the regulator function. Intriguingly, in some scenarios, we find simultaneous points of minimal sensitivity relative to both the regulator and gauge parameters, which allow for a UV completion. This provides further indications that the simultaneous UV-completion of quantum gravity and matter via an asymptotically safe fixed point is a robust physical feature, and that physical quantities, like scaling exponents, can become independent of unphysical choices.

hep-th↗

arXiv2025

Correct Estimation of Higher-Order Spectra: From Theoretical Challenges to Practical Multi-Channel Implementation in SignalSnap

Higher-order spectra (Brillinger's polyspectra) offer powerful methods for solving critical problems in signal processing and data analysis. Despite their significant potential, their practical use has remained limited due to unresolved mathematical issues in spectral estimation, including the absence of unbiased and consistent estimators and the high computational cost associated with evaluating multidimensional spectra. Consequently, existing tools frequently produce artifacts, no existing software library correctly implements Brillinger's cumulant-based trispectrum, or fail to scale effectively to real-world data volumes, leaving crucial applications like multi-detector spectral analysis largely unexplored. In this paper, we revisit higher-order spectra from a modern perspective, addressing the root causes of their historical underuse. We reformulate higher-order spectral estimation using recently derived multivariate k-statistics, yielding unbiased and consistent estimators that eliminate spurious artifacts and precisely align with Brillinger's theoretical definitions. Our methodology covers single- and multi-channel spectral analysis up to the bispectrum (third order) and trispectrum (fourth order), enabling robust investigations of inter-frequency coupling, non-Gaussian behavior, and time-reversal symmetry breaking. Additionally, we introduce quasi-polyspectra to uncover non-stationary, time-dependent higher-order features. We implement these new estimators in SignalSnap, an open-source GPU-accelerated library capable of efficiently analyzing datasets exceeding hundreds of gigabytes within minutes. In applications such as continuous quantum measurements, SignalSnap's rigorous estimators enable precise quantitative matching between experimental data and theoretical models.

physics.data-an↗

arXiv2025

BigReg: An Efficient Registration Pipeline for High-Resolution X-Ray and Light-Sheet Fluorescence Microscopy

Recently, X-ray microscopy (XRM) and light-sheet fluorescence microscopy (LSFM) have emerged as pivotal tools in preclinical research, particularly for studying bone remodeling diseases such as osteoporosis. These modalities offer micrometer-level resolution, and their integration allows for a complementary examination of bone microstructures which is essential for analyzing functional changes. However, registering high-resolution volumes from these independently scanned modalities poses substantial challenges, especially in real-world and reference-free scenarios. This paper presents BigReg, a fast, two-stage pipeline designed for large-volume registration of XRM and LSFM data. The first stage involves extracting surface features and applying two successive point cloud-based methods for coarse alignment. The subsequent stage refines this alignment using a modified cross-correlation technique, achieving precise volumetric registration. Evaluations using expert-annotated landmarks and augmented test data demonstrate that BigReg approaches the accuracy of landmark-based registration with a landmark distance (LMD) of 8.36\,\textmu m\,$\pm$\,0.12\,\textmu m and a landmark fitness (LM fitness) of 85.71\%\,$\pm$\,1.02\%. Moreover, BigReg can provide an optimal initialization for mutual information-based methods which otherwise fail independently, further reducing LMD to 7.24\,\textmu m\,$\pm$\,0.11\,\textmu m and increasing LM fitness to 93.90\%\,$\pm$\,0.77\%. Ultimately, key microstructures, notably lacunae in XRM and bone cells in LSFM, are accurately aligned, enabling unprecedented insights into the pathology of osteoporosis.

cs.CV↗

arXiv2025

CyclePose -- Leveraging Cycle-Consistency for Annotation-Free Nuclei Segmentation in Fluorescence Microscopy

In recent years, numerous neural network architectures specifically designed for the instance segmentation of nuclei in microscopic images have been released. These models embed nuclei-specific priors to outperform generic architectures like U-Nets; however, they require large annotated datasets, which are often not available. Generative models (GANs, diffusion models) have been used to compensate for this by synthesizing training data. These two-stage approaches are computationally expensive, as first a generative model and then a segmentation model has to be trained. We propose CyclePose, a hybrid framework integrating synthetic data generation and segmentation training. CyclePose builds on a CycleGAN architecture, which allows unpaired translation between microscopy images and segmentation masks. We embed a segmentation model into CycleGAN and leverage a cycle consistency loss for self-supervision. Without annotated data, CyclePose outperforms other weakly or unsupervised methods on two public datasets. Code is available at https://github.com/jonasutz/CyclePose

cs.CV↗

arXiv2025

DiffRenderGAN: Addressing Training Data Scarcity in Deep Segmentation Networks for Quantitative Nanomaterial Analysis through Differentiable Rendering and Generative Modelling

Nanomaterials exhibit distinctive properties governed by parameters such as size, shape, and surface characteristics, which critically influence their applications and interactions across technological, biological, and environmental contexts. Accurate quantification and understanding of these materials are essential for advancing research and innovation. In this regard, deep learning segmentation networks have emerged as powerful tools that enable automated insights and replace subjective methods with precise quantitative analysis. However, their efficacy depends on representative annotated datasets, which are challenging to obtain due to the costly imaging of nanoparticles and the labor-intensive nature of manual annotations. To overcome these limitations, we introduce DiffRenderGAN, a novel generative model designed to produce annotated synthetic data. By integrating a differentiable renderer into a Generative Adversarial Network (GAN) framework, DiffRenderGAN optimizes textural rendering parameters to generate realistic, annotated nanoparticle images from non-annotated real microscopy images. This approach reduces the need for manual intervention and enhances segmentation performance compared to existing synthetic data methods by generating diverse and realistic data. Tested on multiple ion and electron microscopy cases, including titanium dioxide (TiO$_2$), silicon dioxide (SiO$_2$)), and silver nanowires (AgNW), DiffRenderGAN bridges the gap between synthetic and real data, advancing the quantification and understanding of complex nanomaterial systems.

cond-mat.mtrl-sci↗

arXiv2024

Spin couplings as witnesses of Planck scale phenomenology

Modified dispersion relations (MDRs) and noncommutative geometries are phenomenological models of Planck-scale corrections to relativistic kinematics, motivated by several approaches to quantum gravity. High-energy astrophysical observations, while commonly used to test such effects, are limited by significant systematic uncertainties. In contrast, low-energy, nonrelativistic experiments provide greater control, with precision serving as an amplifier for Planck-suppressed corrections. We derive corrections to Pauli's equation for nonrelativistic spin-1/2 particles in a magnetic field, incorporating general MDRs and noncommutative geometries. Applying our framework to k-Poincaré symmetries and minimal-length quantum mechanics, we identify Planck-scale corrections accessible in the nonrelativistic regime. Using the electron's anomalous magnetic moment, we constrain model parameters, pushing the k-Poincaré scale in the bi-crossproduct representation beyond $10^{10}$ GeV. These results highlight the complementarity of low-energy precision tests and astrophysical observation in probing quantum gravity phenomenology.

hep-th↗

arXiv2024

Weak equivalence principle and nonrelativistic limit of general dispersion relations

We study the weak equivalence principle in the context of modified dispersion relations, a prevalent approach to quantum gravity phenomenology. We find that generic modified dispersion relations violate the weak equivalence principle. The acceleration in general depends on the mass of the test body, unless the Hamiltonian is either two-homogeneous in the test particles' 4-momenta or the corresponding Lagrangian differs from the homogeneous case by a total derivative only. The key ingredient of this calculation is a $3+1$ decomposition of the parametrization invariant relativistic test particle action derived from the dispersion relation. Additionally, we apply a perturbative expansion in the test particle's spatial velocity and the inverse speed of light. To quantify our result, we provide a general formula for the Eötvós factor of modified dispersion relations. As a specific example, we study the point-particle motion determined from the $κ$-Poincaré dispersion relation in the bicrossproduct basis. Comparing the ensuing non-vanishing Eötvós factor to recent data from the MICROSCOPE experiment, we obtain a bound of the model parameter $\hatΞ {}^{-1}\geq10^{15}{\rm GeV}/c^2$.

gr-qc↗

arXiv2024

A gradient-based approach to fast and accurate head motion compensation in cone-beam CT

Cone-beam computed tomography (CBCT) systems, with their flexibility, present a promising avenue for direct point-of-care medical imaging, particularly in critical scenarios such as acute stroke assessment. However, the integration of CBCT into clinical workflows faces challenges, primarily linked to long scan duration resulting in patient motion during scanning and leading to image quality degradation in the reconstructed volumes. This paper introduces a novel approach to CBCT motion estimation using a gradient-based optimization algorithm, which leverages generalized derivatives of the backprojection operator for cone-beam CT geometries. Building on that, a fully differentiable target function is formulated which grades the quality of the current motion estimate in reconstruction space. We drastically accelerate motion estimation yielding a 19-fold speed-up compared to existing methods. Additionally, we investigate the architecture of networks used for quality metric regression and propose predicting voxel-wise quality maps, favoring autoencoder-like architectures over contracting ones. This modification improves gradient flow, leading to more accurate motion estimation. The presented method is evaluated through realistic experiments on head anatomy. It achieves a reduction in reprojection error from an initial average of 3mm to 0.61mm after motion compensation and consistently demonstrates superior performance compared to existing approaches. The analytic Jacobian for the backprojection operation, which is at the core of the proposed method, is made publicly available. In summary, this paper contributes to the advancement of CBCT integration into clinical workflows by proposing a robust motion estimation approach that enhances efficiency and accuracy, addressing critical challenges in time-sensitive scenarios.

eess.IV↗

arXiv2024

On the Influence of Smoothness Constraints in Computed Tomography Motion Compensation

Computed tomography (CT) relies on precise patient immobilization during image acquisition. Nevertheless, motion artifacts in the reconstructed images can persist. Motion compensation methods aim to correct such artifacts post-acquisition, often incorporating temporal smoothness constraints on the estimated motion patterns. This study analyzes the influence of a spline-based motion model within an existing rigid motion compensation algorithm for cone-beam CT on the recoverable motion frequencies. Results demonstrate that the choice of motion model crucially influences recoverable frequencies. The optimization-based motion compensation algorithm is able to accurately fit the spline nodes for frequencies almost up to the node-dependent theoretical limit according to the Nyquist-Shannon theorem. Notably, a higher node count does not compromise reconstruction performance for slow motion patterns, but can extend the range of recoverable high frequencies for the investigated algorithm. Eventually, the optimal motion model is dependent on the imaged anatomy, clinical use case, and scanning protocol and should be tailored carefully to the expected motion frequency spectrum to ensure accurate motion compensation.

eess.IV↗

arXiv2024

Differentiable Score-Based Likelihoods: Learning CT Motion Compensation From Clean Images

Motion artifacts can compromise the diagnostic value of computed tomography (CT) images. Motion correction approaches require a per-scan estimation of patient-specific motion patterns. In this work, we train a score-based model to act as a probability density estimator for clean head CT images. Given the trained model, we quantify the deviation of a given motion-affected CT image from the ideal distribution through likelihood computation. We demonstrate that the likelihood can be utilized as a surrogate metric for motion artifact severity in the CT image facilitating the application of an iterative, gradient-based motion compensation algorithm. By optimizing the underlying motion parameters to maximize likelihood, our method effectively reduces motion artifacts, bringing the image closer to the distribution of motion-free scans. Our approach achieves comparable performance to state-of-the-art methods while eliminating the need for a representative data set of motion-affected samples. This is particularly advantageous in real-world applications, where patient motion patterns may exhibit unforeseen variability, ensuring robustness without implicit assumptions about recoverable motion types.

cs.CV↗

arXiv2024

Segmentation-Guided Knee Radiograph Generation using Conditional Diffusion Models

Deep learning-based medical image processing algorithms require representative data during development. In particular, surgical data might be difficult to obtain, and high-quality public datasets are limited. To overcome this limitation and augment datasets, a widely adopted solution is the generation of synthetic images. In this work, we employ conditional diffusion models to generate knee radiographs from contour and bone segmentations. Remarkably, two distinct strategies are presented by incorporating the segmentation as a condition into the sampling and training process, namely, conditional sampling and conditional training. The results demonstrate that both methods can generate realistic images while adhering to the conditioning segmentation. The conditional training method outperforms the conditional sampling method and the conventional U-Net.

eess.IV↗

arXiv2024

Physics-Informed Learning for Time-Resolved Angiographic Contrast Agent Concentration Reconstruction

Three-dimensional Digital Subtraction Angiography (3D-DSA) is a well-established X-ray-based technique for visualizing vascular anatomy. Recently, four-dimensional DSA (4D-DSA) reconstruction algorithms have been developed to enable the visualization of volumetric contrast flow dynamics through time-series of volumes. . This reconstruction problem is ill-posed mainly due to vessel overlap in the projection direction and geometric vessel foreshortening, which leads to information loss in the recorded projection images. However, knowledge about the underlying fluid dynamics can be leveraged to constrain the solution space. In our work, we implicitly include this information in a neural network-based model that is trained on a dataset of image-based blood flow simulations. The model predicts the spatially averaged contrast agent concentration for each centerline point of the vasculature over time, lowering the overall computational demand. The trained network enables the reconstruction of relative contrast agent concentrations with a mean absolute error of 0.02 $\pm$ 0.02 and a mean absolute percentage error of 5.31 % $\pm$ 9.25 %. Moreover, the network is robust to varying degrees of vessel overlap and vessel foreshortening. Our approach demonstrates the potential of the integration of machine learning and blood flow simulations in time-resolved angiographic flow reconstruction.

cs.CV↗

arXiv2024

Exploring Epipolar Consistency Conditions for Rigid Motion Compensation in In-vivo X-ray Microscopy

Intravital X-ray microscopy (XRM) in preclinical mouse models is of vital importance for the identification of microscopic structural pathological changes in the bone which are characteristic of osteoporosis. The complexity of this method stems from the requirement for high-quality 3D reconstructions of the murine bones. However, respiratory motion and muscle relaxation lead to inconsistencies in the projection data which result in artifacts in uncompensated reconstructions. Motion compensation using epipolar consistency conditions (ECC) has previously shown good performance in clinical CT settings. Here, we explore whether such algorithms are suitable for correcting motion-corrupted XRM data. Different rigid motion patterns are simulated and the quality of the motion-compensated reconstructions is assessed. The method is able to restore microscopic features for out-of-plane motion, but artifacts remain for more realistic motion patterns including all six degrees of freedom of rigid motion. Therefore, ECC is valuable for the initial alignment of the projection data followed by further fine-tuning of motion parameters using a reconstruction-based method.

cs.CV↗

arXiv2024

Towards quantum Mechanics on the curved cotangent bundle

The minimal-length paradigm is a cornerstone of quantum gravity phenomenology. Recently, it has been demonstrated that minimal-length quantum mechanics can alternatively be described as an undeformed theory set on a nontrivial momentum space. However, there is no fully consistent formulation of these theories beyond Cartesian coordinates in flat space and, in particular, no position representation. This paper is intended to take the first steps in bridging this gap. We find a natural position representation of the position and momentum operators on general curved cotangent bundles. In an expansion akin to Riemann normal coordinates with curvature in both position and momentum space, we apply the formalism perturbatively to the isotropic harmonic oscillator and the hydrogenic atom. Due to the symmetry of the harmonic oscillator under exchange of positions and momenta, we show that it is impossible to distinguish position- from momentum- space curvature with oscillators alone. Thus, we obtain an instantiation of Born reciprocity on the curved cotangent bundle, i. e. in precisely the way Born originally envisioned. It manifests itself as a symmetry mixing UV and IR physics, reminiscent of T-duality in string theory.

gr-qc↗

arXiv2023

Momentum gauge fields from curved momentum space through Kaluza-Klein reduction

In this work we investigate the relation between curved momentum space and momentum-dependent gauge fields. While the former is a classic idea that has been shown to be tied to minimal-length models, the latter constitutes a relatively recent development in quantum gravity phenomenology. In particular, the gauge principle in momentum space amounts to a modification of the position operator of the form $\hat{X}^μ\rightarrow\hat{X}^μ-g A^μ(\hat{P})$ akin to a gauge-covariant derivative in momentum space according to the minimal coupling prescription. Here, we derive both effects from a Kaluza-Klein reduction of a higher-dimensional geometry exhibiting curvature in momentum space. The interplay of the emerging gauge fields as well as the remaining curved momentum space lead to modifications of the Heisenberg algebra. While the gauge fields imply Moyal-type noncommutativity dependent on the analogue field strength tensor, the dimensionally reduced curved momentum space geometry translates to a Snyder-type noncommutative geometry.

gr-qc↗