arXiv subjects

Lukas Hahn

Publications and source records attributed to Lukas Hahn.

10 recordsLinked to original sources

arXiv2026

Texture-Shape Bias Balancing for Robust Synthetic-to-Real Semantic Segmentation in Automotive NIR Imagery

Semantic segmentation is a fundamental component of visual perception in modern automotive systems, enabling pixel-level scene understanding. Near-Infrared imaging (NIR) offers stable detection under difficult illumination conditions, but the development of domain-specific semantic segmentation models remains challenging due to the lack of high-quality annotated data from real-world scenarios. Synthetic datasets offer a scalable alternative, but models trained on synthetic images often suffer performance degradation when transferred to real domains. We present the first systematic study on synthetic to real domain adaptation for semantic segmentation in NIR images in the automotive domain. We propose a generative augmentation framework that transforms synthetic images into realistic NIR-style variants via our introduced target style adaptation (TSA). TSA fine-tunes a latent diffusion model via low-rank adaptation on a small curated set of real NIR images and applies it to synthetic training data using structure-preserving multi-signal conditioning. To reduce texture bias and improve segmentation robustness, we further apply a Voronoi-based style diversification strategy (VSD) that modifies the original textures while preserving scene geometry. Experiments with multiple model architectures on NIR data from vehicle interiors and street scenes show that balancing inductive bias during training leads to noticeably more robust semantic segmentation and effectively reduces the domain gap in our real-world scenarios by up to 63.6% on exterior and 28.4% on interior data. The code is available at GitHub.

cs.CV↗

arXiv2026

InCaRPose: In-Cabin Relative Camera Pose Estimation Model and Dataset

Camera extrinsic calibration is a fundamental task in computer vision. However, precise relative pose estimation in constrained, highly distorted environments, such as in-cabin automotive monitoring (ICAM), remains challenging. We present InCaRPose, a Transformer-based architecture designed for robust relative pose prediction between image pairs, which can be used for camera extrinsic calibration. By leveraging frozen backbone features such as DINOv3 and a Transformer-based decoder, our model effectively captures the geometric relationship between a reference and a target view. Unlike traditional methods, our approach achieves absolute metric-scale translation within the physically plausible adjustment range of in-cabin camera mounts in a single inference step, which is critical for ICAM, where accurate real-world distances are required for safety-relevant perception. We specifically address the challenges of highly distorted fisheye cameras in automotive interiors by training exclusively on synthetic data. Our model is capable of generalization to real-world cabin environments without relying on the exact same camera intrinsics and additionally achieves competitive performance on the public 7-Scenes dataset. Despite having limited training data, InCaRPose maintains high precision in both rotation and translation, even with a ViT-Small backbone. This enables real-time performance for time-critical inference, such as driver monitoring in supervised autonomous driving. We release our real-world In-Cabin-Pose test dataset consisting of highly distorted vehicle-interior images and our code at https://github.com/felixstillger/InCaRPose.

cs.CV↗

arXiv2026

Mathematical Foundations of Modeling ETL Process Chains

Extract-Transform-Load (ETL) processes are core components of modern data processing infrastructures. The throughput of processed data records can be adjusted by changing the amount of allocated resources, i.e.~the number of parallel processing threads for each of the three ETL phases, but also depends on stochastic variations in the per-record processing times. In chains of multiple consecutive ETL processes, the relation between allocated resources and overall throughput is further complicated, for example by the occurrence of bottlenecks affecting all subsequent ETL processes. We develop a mathematical model of ETL process chains that is accurate at the level of time-aggregated throughput and suitable for efficient simulation. The process chain is represented as a controlled discrete-time Markov process on a directed acyclic graph whose edges are individual ETL processes. We model the mean throughput as a bounded, monotone function of the number of parallel threads, to capture the diminishing benefit of allocating more threads. We furthermore introduce a Flow Balance postulate linking number of threads, mean throughput, and mean processing time. The stochastic processing times are then modeled by non-negative heavy-tailed distributions around the mean processing time. This framework provides a principled simulator for ETL networks and a foundation for learning- and control-based resource allocation.

cs.DC↗

arXiv2022

MuRiT: Efficient Computation of Pathwise Persistence Barcodes in Multi-Filtered Flag Complexes via Vietoris-Rips Transformations

Multi-parameter persistent homology naturally arises in applications of persistent topology to data that come with extra information depending on additional parameters, like for example time series data. We introduce the concept of a Vietoris-Rips transformation, a method that reduces the computation of the one-parameter persistent homology of pathwise subcomplexes in multi-filtered flag complexes to the computation of the Vietoris-Rips persistent homology of certain semimetric spaces. The corresponding pathwise persistence barcodes track persistence features of the ambient multi-filtered complex and can in particular be used to recover the rank invariant in multi-parameter persistent homology. We present MuRiT, a scalable algorithm that computes the pathwise persistence barcodes of multi-filtered flag complexes by means of Vietoris-Rips transformations. Moreover, we provide an efficient software implementation of the MuRiT algorithm which resorts to Ripser for the actual computation of Vietoris-Rips persistence barcodes. To demonstrate the applicability of MuRiT to real-world datasets, we establish MuRiT as part of our CoVtRec pipeline for the surveillance of the convergent evolution of the coronavirus SARS-CoV-2 in the current COVID-19 pandemic.

math.AT↗

arXiv2022

A-model Implications of Extended Mirror Symmetry

Associativity of the quantum product ensures flatness of the Dubrovin connection and is the basis for Hodge-theoretic mirror symmetry of Calabi-Yau threefolds. We use ring and module structure on cohomology pertaining to a Lagrangian submanifold to define an extension of the A-model Variation of mixed Hodge structure that matches the predictions from extended mirror symmetry. Our construction makes contact with axioms for open Gromov-Witten theory recently proposed by Solomon-Tukachinsky.

math.AG↗

arXiv2021

Ultrafast topological data analysis reveals pandemic-scale dynamics of convergent evolution

Genome variants which re-occur independently across evolutionary lineages are key molecular signatures of adaptation. Inferring the dynamics of such genetic changes from pandemic-scale genomic datasets is now possible, which opens up unprecedented insight into evolutionary processes. However, existing approaches depend on the construction of accurate phylogenetic trees, which remains challenging at scale. Here we present EVOtRec, an organism-agnostic, fast and scalable Topological Data Analysis approach that enables the inference of convergently evolving genomic variants over time directly from topological patterns in the dataset, without requiring the construction of a phylogenetic tree. Using data from both simulations and published experiments, we show that EVOtRec can robustly identify variants under positive selection and performs orders of magnitude faster than state-of-the-art phylogeny-based approaches, with comparable results. We apply EVOtRec to three large viral genome datasets: SARS-CoV-2, influenza virus A subtype H5N1 and HIV-1. We identify key convergent genome variants and demonstrate how EVOtRec facilitates the real-time tracking of high fitness variants in large datasets with millions of genomes, including effects modulated by varying genomic backgrounds. We envision our Topological Data Analysis approach as a new framework for efficient comparative genomics.

q-bio.PE↗

arXiv2020

FLIC: Fast Lidar Image Clustering

Lidar sensors are widely used in various applications, ranging from scientific fields over industrial use to integration in consumer products. With an ever growing number of different driver assistance systems, they have been introduced to automotive series production in recent years and are considered an important building block for the practical realisation of autonomous driving. However, due to the potentially large amount of Lidar points per scan, tailored algorithms are required to identify objects (e.g. pedestrians or vehicles) with high precision in a very short time. In this work, we propose an algorithmic approach for real-time instance segmentation of Lidar sensor data. We show how our method leverages the properties of the Euclidean distance to retain three-dimensional measurement information, while being narrowed down to a two-dimensional representation for fast computation. We further introduce what we call "skip connections", to make our approach robust against over-segmentation and improve assignment in cases of partial occlusion. Through detailed evaluation on public data and comparison with established methods, we show how these aspects enable state-of-the-art performance and runtime on a single CPU core.

cs.CV↗

arXiv2020

On the Robustness of Active Learning

Active Learning is concerned with the question of how to identify the most useful samples for a Machine Learning algorithm to be trained with. When applied correctly, it can be a very powerful tool to counteract the immense data requirements of Artificial Neural Networks. However, we find that it is often applied with not enough care and domain knowledge. As a consequence, unrealistic hopes are raised and transfer of the experimental results from one dataset to another becomes unnecessarily hard. In this work we analyse the robustness of different Active Learning methods with respect to classifier capacity, exchangeability and type, as well as hyperparameters and falsely labelled data. Experiments reveal possible biases towards the architecture used for sample selection, resulting in suboptimal performance for other classifiers. We further propose the new "Sum of Squared Logits" method based on the Simpson diversity index and investigate the effect of using the confusion matrix for balancing in sample selection.

cs.CV↗

arXiv2020

Fast Object Classification and Meaningful Data Representation of Segmented Lidar Instances

Object detection algorithms for Lidar data have seen numerous publications in recent years, reporting good results on dataset benchmarks oriented towards automotive requirements. Nevertheless, many of these are not deployable to embedded vehicle systems, as they require immense computational power to be executed close to real time. In this work, we propose a way to facilitate real-time Lidar object classification on CPU. We show how our approach uses segmented object instances to extract important features, enabling a computationally efficient batch-wise classification. For this, we introduce a data representation which translates three-dimensional information into small image patches, using decomposed normal vector images. We couple this with dedicated object statistics to handle edge cases. We apply our method on the tasks of object detection and semantic segmentation, as well as the relatively new challenge of panoptic segmentation. Through evaluation, we show, that our algorithm is capable of producing good results on public data, while running in real time on CPU without using specific optimisation.

cs.CV↗

arXiv2019

Fast and Reliable Architecture Selection for Convolutional Neural Networks

The performance of a Convolutional Neural Network (CNN) depends on its hyperparameters, like the number of layers, kernel sizes, or the learning rate for example. Especially in smaller networks and applications with limited computational resources, optimisation is key. We present a fast and efficient approach for CNN architecture selection. Taking into account time consumption, precision and robustness, we develop a heuristic to quickly and reliably assess a network's performance. In combination with Bayesian optimisation (BO), to effectively cover the vast parameter space, our contribution offers a plain and powerful architecture search for this machine learning technique.

cs.CV↗