arXiv subjects

Clemens Heitzinger

Publications and source records attributed to Clemens Heitzinger.

14 recordsLinked to original sources

arXiv2025

One-Class Domain Adaptation via Meta-Learning

The deployment of IoT (Internet of Things) sensor-based machine learning models in industrial systems for anomaly classification tasks poses significant challenges due to distribution shifts, as the training data acquired in controlled laboratory settings may significantly differ from real-time data in production environments. Furthermore, many real-world applications cannot provide a substantial number of labeled examples for each anomalous class in every new environment. It is therefore crucial to develop adaptable machine learning models that can be effectively transferred from one environment to another, enabling rapid adaptation using normal operational data. We extended this problem setting to an arbitrary classification task and formulated the one-class domain adaptation (OC-DA) problem setting. We took a meta-learning approach to tackle the challenge of OC-DA, and proposed a task sampling strategy to adapt any bi-level meta-learning algorithm to OC-DA. We modified the well-established model-agnostic meta-learning (MAML) algorithm and introduced the OC-DA MAML algorithm. We provided a theoretical analysis showing that OC-DA MAML optimizes for meta-parameters that enable rapid one-class adaptation across domains. The OC-DA MAML algorithm is evaluated on the Rainbow-MNIST meta-learning benchmark and on a real-world dataset of vibration-based sensor readings. The results show that OC-DA MAML significantly improves the performance on the target domains and outperforms MAML using the standard task sampling strategy.

cs.LG↗

arXiv2025

Towards Fingerprint Mosaicking Artifact Detection: A Self-Supervised Deep Learning Approach

Fingerprint mosaicking, which is the process of combining multiple fingerprint images into a single master fingerprint, is an essential process in modern biometric systems. However, it is prone to errors that can significantly degrade fingerprint image quality. This paper proposes a novel deep learning-based approach to detect and score mosaicking artifacts in fingerprint images. Our method leverages a self-supervised learning framework to train a model on large-scale unlabeled fingerprint data, eliminating the need for manual artifact annotation. The proposed model effectively identifies mosaicking errors, achieving high accuracy on various fingerprint modalities, including contactless, rolled, and pressed fingerprints and furthermore proves to be robust to different data sources. Additionally, we introduce a novel mosaicking artifact score to quantify the severity of errors, enabling automated evaluation of fingerprint images. By addressing the challenges of mosaicking artifact detection, our work contributes to improving the accuracy and reliability of fingerprint-based biometric systems.

cs.CV↗

arXiv2025

TipSegNet: Fingertip Segmentation in Contactless Fingerprint Imaging

Contactless fingerprint recognition systems offer a hygienic, user-friendly, and efficient alternative to traditional contact-based methods. However, their accuracy heavily relies on precise fingertip detection and segmentation, particularly under challenging background conditions. This paper introduces TipSegNet, a novel deep learning model that achieves state-of-the-art performance in segmenting fingertips directly from grayscale hand images. TipSegNet leverages a ResNeXt-101 backbone for robust feature extraction, combined with a Feature Pyramid Network (FPN) for multi-scale representation, enabling accurate segmentation across varying finger poses and image qualities. Furthermore, we employ an extensive data augmentation strategy to enhance the model's generalizability and robustness. TipSegNet outperforms existing methods, achieving a mean Intersection over Union (mIoU) of 0.987 and an accuracy of 0.999, representing a significant advancement in contactless fingerprint segmentation. This enhanced accuracy has the potential to substantially improve the reliability and effectiveness of contactless biometric systems in real-world applications.

cs.CV↗

arXiv2023

Centrality of the Fingerprint Core Location

Fingerprints have long been recognized as a unique and reliable means of personal identification. Central to the analysis and enhancement of fingerprints is the concept of the fingerprint core. Although the location of the core is used in many applications, to the best of our knowledge, this study is the first to investigate the empirical distribution of the core over a large, combined dataset of rolled, as well as plain fingerprint recordings. We identify and investigate the extent of incomplete rolling during the rolled fingerprint acquisition and investigate the centrality of the core. After correcting for the incomplete rolling, we find that the core deviates from the fingerprint center by 5.7% $\pm$ 5.2% to 7.6% $\pm$ 6.9%, depending on the finger. Additionally, we find that the assumption of normal distribution of the core position of plain fingerprint recordings cannot be rejected, but for rolled ones it can. Therefore, we use a multi-step process to find the distribution of the rolled fingerprint recordings. The process consists of an Anderson-Darling normality test, the Bayesian Information Criterion to reduce the number of possible candidate distributions and finally a Generalized Monte Carlo goodness-of-fit procedure to find the best fitting distribution. We find the non-central Fischer distribution best describes the cores' horizontal positions. Finally, we investigate the correlation between mean core position offset and the NFIQ 2 score and find that the NFIQ 2 prefers rolled fingerprint recordings where the core sits slightly below the fingerprint center.

cs.CV↗

arXiv2022

Autoencoder based Anomaly Detection and Explained Fault Localization in Industrial Cooling Systems

Anomaly detection in large industrial cooling systems is very challenging due to the high data dimensionality, inconsistent sensor recordings, and lack of labels. The state of the art for automated anomaly detection in these systems typically relies on expert knowledge and thresholds. However, data is viewed isolated and complex, multivariate relationships are neglected. In this work, we present an autoencoder based end-to-end workflow for anomaly detection suitable for multivariate time series data in large industrial cooling systems, including explained fault localization and root cause analysis based on expert knowledge. We identify system failures using a threshold on the total reconstruction error (autoencoder reconstruction error including all sensor signals). For fault localization, we compute the individual reconstruction error (autoencoder reconstruction error for each sensor signal) allowing us to identify the signals that contribute most to the total reconstruction error. Expert knowledge is provided via look-up table enabling root-cause analysis and assignment to the affected subsystem. We demonstrated our findings in a cooling system unit including 34 sensors over a 8-months time period using 4-fold cross validation approaches and automatically created labels based on thresholds provided by domain experts. Using 4-fold cross validation, we reached a F1-score of 0.56, whereas the autoencoder results showed a higher consistency score (CS of 0.92) compared to the automatically created labels (CS of 0.62) -- indicating that the anomaly is recognized in a very stable manner. The main anomaly was found by the autoencoder and automatically created labels and was also recorded in the log files. Further, the explained fault localization highlighted the most affected component for the main anomaly in a very consistent manner.

cs.LG↗

arXiv2021

Evaluation of Hyperparameter-Optimization Approaches in an Industrial Federated Learning System

Federated Learning (FL) decouples model training from the need for direct access to the data and allows organizations to collaborate with industry partners to reach a satisfying level of performance without sharing vulnerable business information. The performance of a machine learning algorithm is highly sensitive to the choice of its hyperparameters. In an FL setting, hyperparameter optimization poses new challenges. In this work, we investigated the impact of different hyperparameter optimization approaches in an FL system. In an effort to reduce communication costs, a critical bottleneck in FL, we investigated a local hyperparameter optimization approach that -- in contrast to a global hyperparameter optimization approach -- allows every client to have its own hyperparameter configuration. We implemented these approaches based on grid search and Bayesian optimization and evaluated the algorithms on the MNIST data set using an i.i.d. partition and on an Internet of Things (IoT) sensor based industrial data set using a non-i.i.d. partition.

cs.LG↗

arXiv2020

Frequency dependence of dielectrophoresis fabrication of single-walled carbon nanotube field-effect transistor

In this paper, we present a new theoretical model for the dielectrophoresis (DEP) process of the single-walled carbon nanotubes (SWCNT). We obtain a different frequency interval for the alignment of wide energy gap semiconductor SWCNTs which shows a considerable difference with the prevalent model. For this, we study two specific models namely, the spherical model and the ellipsoid model to estimate the frequency interval. Then, we perform the DEP process and use the obtained frequencies (of spherical and ellipsoid models) for the alignment of the SWCNTs. Our empirical results declare the theoretical prediction, i.e., a crucial step toward the realization of carbon nanotube field-effect transistor (CNT-FET) with the DEP process based on the ellipsoid model.

physics.app-ph↗

arXiv2019

Stochastic Galerkin Methods for the Boltzmann-Poisson system

We study uncertainty quantification for a Boltzmann-Poisson system that models electron transport in semiconductors and the physical collision mechanisms over the charges. We use the stochastic Galerkin method in order to handle the randomness associated with the problem. The main uncertainty in the Boltzmann equation concerns the initial conditions for a large number of particles, which is why the problem is formulated in terms of a probability density in phase space. The second source of uncertainty, directly related to the quantum nature of the problem, is the collision operator, as its structure in this semiclassical model comes from the quantum scattering matrices operating on the wave function associated to the electron probability density. Additional sources of uncertainty are transport, boundary data, etc. In this study we choose first the phonon energy as a random variable, since its value influences the energy jump appearing in the collision integral for electron-phonon scattering. Then we choose the lattice temperature as a random variable, since it defines the value of the collision operator terms in the case of electron-phonon scattering by being a parameter of the phonon distribution. The random variable for this case is a scalar then. Finally, we present our numerical simulations.

math.NA↗

arXiv2019

Bayesian inversion for nanowire field-effect sensors

Nanowire field-effect sensors have recently been developed for label-free detection of biomolecules. In this work, we introduce a computational technique based on Bayesian estimation to determine the physical parameters of the sensor and, more importantly, the properties of the analyte molecules. To that end, we first propose a PDE based model to simulate the device charge transport and electrochemical behavior. Then, the adaptive Metropolis algorithm with delayed rejection (DRAM) is applied to estimate the posterior distribution of unknown parameters, namely molecule charge density, molecule density, doping concentration, and electron and hole mobilities. We determine the device and molecules properties simultaneously, and we also calculate the molecule density as the only parameter after having determined the device parameters. This approach makes it possible not only to determine unknown parameters, but it also shows how well each parameter can be determined by yielding the probability density function (pdf).

math.NA↗

arXiv2019

A Bayesian estimation method for variational phase-field fracture problems

In this work, we propose a parameter estimation framework for fracture propagation problems. The fracture problem is described by a phase-field method. Parameter estimation is realized with a Bayesian framework. Here, the focus is on uncertainties arising in the solid material parameters and the critical energy release rate. A reference value (obtained on a sufficiently small mesh) as the replacement of measurement will be chosen, and their posterior distribution is obtained. Due to time- and mesh dependency of the problem, the computational costs can be high. Using Bayesian inversion, we solve the problem on a relatively coarse mesh and fit the parameters. The obtained load-displacement curve that is usually the target function is matched with the reference values. Finally, our algorithmic approach is substantiated with several numerical examples.

math.NA↗

arXiv2019

An adaptive multilevel Monte Carlo algorithm for the stochastic drift-diffusion-Poisson system

We present an adaptive multilevel Monte Carlo algorithm for solving the stochastic drift-diffusion-Poisson system with non-zero recombination rate. The a-posteriori error is estimated to enable goal-oriented adaptive mesh refinement for the spatial dimensions, while the a-priori error is estimated to guarantee \red{linear} convergence of the $H^1$ error. In the adaptive mesh refinement, efficient estimation of the error indicator gives rise to better error control. For the stochastic dimensions, we use the multilevel Monte Carlo method to solve this system of stochastic partial differential equations. Finally, the advantage of the technique developed here compared to uniform mesh refinement is discussed using a realistic numerical example.

math.NA↗

arXiv2019

Efficient computational design and optimization of dielectric metamaterial devices

Dielectric structures composed of many inclusions that manipulate light in ways the bulk materials cannot are commonly seen in the field of metamaterials. In these structures, each inclusion depends on a set of parameters such as location and orientation, which are difficult to ascertain. We propose and implement an optimization-based approach for designing such metamaterials in two dimensions by using a fast boundary element method and a multiple-scattering solver for a given set of parameters. This approach provides the backbone of an automated process for the design and analysis of metamaterials that does not rely on analytical approximations. We demonstrate the validity of our approach with simulations that converge to optimal parameter values and result in substantially better performance.

physics.comp-ph↗

arXiv2016

The optimal multilevel Monte-Carlo approximation of the stochastic drift-diffusion-Poisson system

Existence and local-uniqueness theorems for weak solutions of a system consisting of the drift-diffusion-Poisson equations and the Poisson-Boltzmann equation, all with stochastic coefficients, are presented. For the numerical approximation of the expected value of the solution of the system, we develop a multi-level Monte-Carlo (MLMC) finite-element method (FEM) and we analyze its rate of convergence and its computational complexity. This allows to find the optimal choice of discretization parameters. Finally, numerical results show the efficiency of the method. Applications are, among others, noise and fluctuations in nanoscale transistors, in field-effect bio- and gas sensors, and in nanopores.

math.AP↗

arXiv2016

Adaptive and Iterative Methods for Simulations of Nanopores with the PNP-Stokes Equations

We present a 3D finite element solver for the nonlinear Poisson-Nernst-Planck (PNP) equations for electrodiffusion, coupled to the Stokes system of fluid dynamics. The model serves as a building block for the simulation of macromolecule dynamics inside nanopore sensors. We add to existing numerical approaches by deploying goal-oriented adaptive mesh refinement. To reduce the computation overhead of mesh adaptivity, our error estimator uses the much cheaper Poisson-Boltzmann equation as a simplified model, which is justified on heuristic grounds but shown to work well in practice. To address the nonlinearity in the full PNP-Stokes system, three different linearization schemes are proposed and investigated, with two segregated iterative approaches both outperforming a naive application of Newton's method. Numerical experiments are reported on a real-world nanopore sensor geometry. We also investigate two different models for the interaction of target molecules with the nanopore sensor through the PNP-Stokes equations. In one model, the molecule is of finite size and is explicitly built into the geometry; while in the other, the molecule is located at a single point and only modeled implicitly -- after solution of the system -- which is computationally favorable. We compare the resulting force profiles of the electric and velocity fields acting on the molecule, and conclude that the point-size model fails to capture important physical effects such as the dependence of charge selectivity of the sensor on the molecule radius.

physics.comp-ph↗