arXiv subjects

Alexander Braun

Publications and source records attributed to Alexander Braun.

At least 19 recordsLinked to original sources

arXiv2026

Mask What Matters: Saliency-Guided Video Self-Supervised Learning for Autonomous Driving

Video self-supervised learning through masked spatiotemporal prediction has emerged as a promising paradigm for learning feature representations from unlabeled data. However, existing methods typically rely on random masking, which indiscriminately removes regions irrespective of their semantic or temporal relevance. In ego-centric driving videos, this can weaken the pretext signal since safety-critical cues such as pedestrians, vehicles, lane boundaries, and dynamic interactions often occupy only a small portion of the frame, yet are central to downstream perception. We introduce V-JEPA4A, a domain-specialized variant of V-JEPA for autonomous driving that is pre-trained on publicly available driving videos with a novel saliency-driven masking policy. It accounts for semantically and temporally relevant context. The proposed policy preserves and predicts context according to semantic importance and temporal relevance, yielding more informative representation learning while retaining the efficiency of masked prediction. We evaluate the resulting encoders on four driving benchmarks spanning tracking, semantic segmentation, and depth estimation. The results demonstrate that V-JEPA4A reduces identity switches on BDD100k MOT by 25% over V-JEPA with random masking, achieves 73.2 mIoU on Cityscapes, and 3.75 RMSE on KITTI-2015 depth, while incurring only ~14% additional pre-training iteration overhead.

cs.CV↗

arXiv2026

Analyzing Training-Free Corruption Detection for Object Detection Datasets

Annotation errors are widespread in computer vision datasets and can significantly degrade the performance of systems trained on them, particularly in complex tasks such as object detection. Several approaches exist to identify annotation errors, including training-free feature-space methods which provide a fast and interpretable way to analyze annotations. However, the behavior on object detection annotations, which include semantic and spatial information, remains largely unexplored. In this work we analyze the applicability of feature-space-based approaches for detecting annotation errors in object detection datasets. By adapting an existing feature-space method, we show that such approaches reliably expose semantic mislabel, while positional errors remain difficult to detect. We evaluate this behavior across multiple pretrained embedding models, synthetic noise types (symmetric, asymmetric, and positional), and real-world annotation errors using VOC2012 and KITTI. All code and real-world corruptions are publicly available at the following repository: https://github.com/ ChristianSieberichs/BoundingBox\_corruption\_detection

cs.CV↗

arXiv2025

Examining the Impact of Optical Aberrations to Image Classification and Object Detection Models

Deep neural networks (DNNs) have proven to be successful in various computer vision applications such that models even infer in safety-critical situations. Therefore, vision models have to behave in a robust way to disturbances such as noise or blur. While seminal benchmarks exist to evaluate model robustness to diverse corruptions, blur is often approximated in an overly simplistic way to model defocus, while ignoring the different blur kernel shapes that result from optical systems. To study model robustness against realistic optical blur effects, this paper proposes two datasets of blur corruptions, which we denote OpticsBench and LensCorruptions. OpticsBench examines primary aberrations such as coma, defocus, and astigmatism, i.e. aberrations that can be represented by varying a single parameter of Zernike polynomials. To go beyond the principled but synthetic setting of primary aberrations, LensCorruptions samples linear combinations in the vector space spanned by Zernike polynomials, corresponding to 100 real lenses. Evaluations for image classification and object detection on ImageNet and MSCOCO show that for a variety of different pre-trained models, the performance on OpticsBench and LensCorruptions varies significantly, indicating the need to consider realistic image corruptions to evaluate a model's robustness against blur.

cs.CV↗

arXiv2025

SOLAS: Superpositioning an Optical Lens in Automotive Simulation

Automotive Simulation is a potentially cost-effective strategy to identify and test corner case scenarios in automotive perception. Recent work has shown a significant shift in creating realistic synthetic data for road traffic scenarios using a video graphics engine. However, a gap exists in modeling realistic optical aberrations associated with cameras in automotive simulation. This paper builds on the concept from existing literature to model optical degradations in simulated environments using the Python-based ray-tracing library KrakenOS. As a novel pipeline, we degrade automotive fisheye simulation using an optical doublet with +/-2 deg Field of View (FOV), introducing realistic optical artifacts into two simulation images from SynWoodscape and Parallel Domain Woodscape. We evaluate KrakenOS by calculating the Root Mean Square Error (RMSE), which averaged around 0.023 across the RGB light spectrum compared to Ansys Zemax OpticStudio, an industrial benchmark for optical design and simulation. Lastly, we measure the image sharpness of the degraded simulation using the ISO12233:2023 Slanted Edge Method and show how both qualitative and measured results indicate the extent of the spatial variation in image sharpness from the periphery to the center of the degradations.

eess.SP↗

arXiv2024

Optical aberrations in autonomous driving: Physics-informed parameterized temperature scaling for neural network uncertainty calibration

'A trustworthy representation of uncertainty is desirable and should be considered as a key feature of any machine learning method' (Huellermeier and Waegeman, 2021). This conclusion of Huellermeier et al. underpins the importance of calibrated uncertainties. Since AI-based algorithms are heavily impacted by dataset shifts, the automotive industry needs to safeguard its system against all possible contingencies. One important but often neglected dataset shift is caused by optical aberrations induced by the windshield. For the verification of the perception system performance, requirements on the AI performance need to be translated into optical metrics by a bijective mapping. Given this bijective mapping it is evident that the optical system characteristics add additional information about the magnitude of the dataset shift. As a consequence, we propose to incorporate a physical inductive bias into the neural network calibration architecture to enhance the robustness and the trustworthiness of the AI target application, which we demonstrate by using a semantic segmentation task as an example. By utilizing the Zernike coefficient vector of the optical system as a physical prior we can significantly reduce the mean expected calibration error in case of optical aberrations. As a result, we pave the way for a trustworthy uncertainty representation and for a holistic verification strategy of the perception chain.

cs.CV↗

arXiv2024

The Secretary Problem with Predicted Additive Gap

The secretary problem is one of the fundamental problems in online decision making; a tight competitive ratio for this problem of $1/\mathrm{e} \approx 0.368$ has been known since the 1960s. Much more recently, the study of algorithms with predictions was introduced: The algorithm is equipped with a (possibly erroneous) additional piece of information upfront which can be used to improve the algorithm's performance. Complementing previous work on secretary problems with prior knowledge, we tackle the following question: What is the weakest piece of information that allows us to break the $1/\mathrm{e}$ barrier? To this end, we introduce the secretary problem with predicted additive gap. As in the classical problem, weights are fixed by an adversary and elements appear in random order. In contrast to previous variants of predictions, our algorithm only has access to a much weaker piece of information: an \emph{additive gap} $c$. This gap is the difference between the highest and $k$-th highest weight in the sequence. Unlike previous pieces of advice, knowing an exact additive gap does not make the problem trivial. Our contribution is twofold. First, we show that for any index $k$ and any gap $c$, we can obtain a competitive ratio of $0.4$ when knowing the exact gap (even if we do not know $k$), hence beating the prevalent bound for the classical problem by a constant. Second, a slightly modified version of our algorithm allows to prove standard robustness-consistency properties as well as improved guarantees when knowing a range for the error of the prediction.

cs.DS↗

arXiv2024

SLAM2REF: Advancing Long-Term Mapping with 3D LiDAR and Reference Map Integration for Precise 6-DoF Trajectory Estimation and Map Extension

This paper presents a pioneering solution to the task of integrating mobile 3D LiDAR and inertial measurement unit (IMU) data with existing building information models or point clouds, which is crucial for achieving precise long-term localization and mapping in indoor, GPS-denied environments. Our proposed framework, SLAM2REF, introduces a novel approach for automatic alignment and map extension utilizing reference 3D maps. The methodology is supported by a sophisticated multi-session anchoring technique, which integrates novel descriptors and registration methodologies. Real-world experiments reveal the framework's remarkable robustness and accuracy, surpassing current state-of-the-art methods. Our open-source framework's significance lies in its contribution to resilient map data management, enhancing processes across diverse sectors such as construction site monitoring, emergency response, disaster management, and others, where fast-updated digital 3D maps contribute to better decision-making and productivity. Moreover, it offers advancements in localization and mapping research. Link to the repository: https://github.com/MigVega/SLAM2REF, Data: https://doi.org/10.14459/2024mp1743877.

cs.RO↗

arXiv2024

BIM-SLAM: Integrating BIM Models in Multi-session SLAM for Lifelong Mapping using 3D LiDAR

While 3D LiDAR sensor technology is becoming more advanced and cheaper every day, the growth of digitalization in the AEC industry contributes to the fact that 3D building information models (BIM models) are now available for a large part of the built environment. These two facts open the question of how 3D models can support 3D LiDAR long-term SLAM in indoor, GPS-denied environments. This paper proposes a methodology that leverages BIM models to create an updated map of indoor environments with sequential LiDAR measurements. Session data (pose graph-based map and descriptors) are initially generated from BIM models. Then, real-world data is aligned with the session data from the model using multi-session anchoring while minimizing the drift on the real-world data. Finally, the new elements not present in the BIM model are identified, grouped, and reconstructed in a surface representation, allowing a better visualization next to the BIM model. The framework enables the creation of a coherent map aligned with the BIM model that does not require prior knowledge of the initial pose of the robot, and it does not need to be inside the map.

cs.RO↗

arXiv2024

SS-SFR: Synthetic Scenes Spatial Frequency Response on Virtual KITTI and Degraded Automotive Simulations for Object Detection

Automotive simulation can potentially compensate for a lack of training data in computer vision applications. However, there has been little to no image quality evaluation of automotive simulation and the impact of optical degradations on simulation is little explored. In this work, we investigate Virtual KITTI and the impact of applying variations of Gaussian blur on image sharpness. Furthermore, we consider object detection, a common computer vision application on three different state-of-the-art models, thus allowing us to characterize the relationship between object detection and sharpness. It was found that while image sharpness (MTF50) degrades from an average of 0.245cy/px to approximately 0.119cy/px; object detection performance stays largely robust within 0.58\%(Faster RCNN), 1.45\%(YOLOF) and 1.93\%(DETR) across all respective held-out test sets.

cs.CV↗

arXiv2024

Approximating Optimum Online for Capacitated Resource Allocation

We study online capacitated resource allocation, a natural generalization of online stochastic max-weight bipartite matching. This problem is motivated by ride-sharing and Internet advertising applications, where online arrivals may have the capacity to serve multiple offline users. Our main result is a polynomial-time online algorithm which is $(1/2 + \kappa)$-approximate to the optimal online algorithm for $\kappa = 0.0115$. This can be contrasted to the (tight) $1/2$-competitive algorithms to the optimum offline benchmark from the prophet inequality literature. Optimum online is a recently popular benchmark for online Bayesian problems which can use unbounded computation, but not "prophetic" knowledge of future inputs. Our algorithm (which also works for the case of stochastic rewards) rounds a generalized LP relaxation from the unit-capacity case via a two-proposal algorithm, as in previous works in the online matching literature. A key technical challenge in deriving our guarantee is bounding the positive correlation among users introduced when rounding our LP relaxation online. Unlike in the case of unit capacities, this positive correlation is unavoidable for guarantees beyond $1/2$. Conceptually, our results show that the study of optimum online as a benchmark can reveal problem-specific insights that are irrelevant to competitive analysis.

cs.DS↗

arXiv2024

Decoupling of neural network calibration measures

A lot of effort is currently invested in safeguarding autonomous driving systems, which heavily rely on deep neural networks for computer vision. We investigate the coupling of different neural network calibration measures with a special focus on the Area Under the Sparsification Error curve (AUSE) metric. We elaborate on the well-known inconsistency in determining optimal calibration using the Expected Calibration Error (ECE) and we demonstrate similar issues for the AUSE, the Uncertainty Calibration Score (UCS), as well as the Uncertainty Calibration Error (UCE). We conclude that the current methodologies leave a degree of freedom, which prevents a unique model calibration for the homologation of safety-critical functionalities. Furthermore, we propose the AUSE as an indirect measure for the residual uncertainty, which is irreducible for a fixed network architecture and is driven by the stochasticity in the underlying data generation process (aleatoric contribution) as well as the limitation in the hypothesis space (epistemic contribution).

cs.CV↗

arXiv2024

A Point-Based Approach to Efficient LiDAR Multi-Task Perception

Multi-task networks can potentially improve performance and computational efficiency compared to single-task networks, facilitating online deployment. However, current multi-task architectures in point cloud perception combine multiple task-specific point cloud representations, each requiring a separate feature encoder and making the network structures bulky and slow. We propose PAttFormer, an efficient multi-task architecture for joint semantic segmentation and object detection in point clouds that only relies on a point-based representation. The network builds on transformer-based feature encoders using neighborhood attention and grid-pooling and a query-based detection decoder using a novel 3D deformable-attention detection head design. Unlike other LiDAR-based multi-task architectures, our proposed PAttFormer does not require separate feature encoders for multiple task-specific point cloud representations, resulting in a network that is 3x smaller and 1.4x faster while achieving competitive performance on the nuScenes and KITTI benchmarks for autonomous driving perception. Our extensive evaluations show substantial gains from multi-task learning, improving LiDAR semantic segmentation by +1.7% in mIou and 3D object detection by +1.7% in mAP on the nuScenes benchmark compared to the single-task models.

cs.CV↗

arXiv2023

Classification robustness to common optical aberrations

Computer vision using deep neural networks (DNNs) has brought about seminal changes in people's lives. Applications range from automotive, face recognition in the security industry, to industrial process monitoring. In some cases, DNNs infer even in safety-critical situations. Therefore, for practical applications, DNNs have to behave in a robust way to disturbances such as noise, pixelation, or blur. Blur directly impacts the performance of DNNs, which are often approximated as a disk-shaped kernel to model defocus. However, optics suggests that there are different kernel shapes depending on wavelength and location caused by optical aberrations. In practice, as the optical quality of a lens decreases, such aberrations increase. This paper proposes OpticsBench, a benchmark for investigating robustness to realistic, practically relevant optical blur effects. Each corruption represents an optical aberration (coma, astigmatism, spherical, trefoil) derived from Zernike Polynomials. Experiments on ImageNet show that for a variety of different pre-trained DNNs, the performance varies strongly compared to disk-shaped kernels, indicating the necessity of considering realistic image degradations. In addition, we show on ImageNet-100 with OpticsAugment that robustness can be increased by using optical kernels as data augmentation. Compared to a conventionally trained ResNeXt50, training with OpticsAugment achieves an average performance gain of 21.7% points on OpticsBench and 6.8% points on 2D common corruptions.

cs.CV↗

arXiv2023

Occupancy Grid Map to Pose Graph-based Map: Robust BIM-based 2D-LiDAR Localization for Lifelong Indoor Navigation in Changing and Dynamic Environments

Several studies rely on the de facto standard Adaptive Monte Carlo Localization (AMCL) method to localize a robot in an Occupancy Grid Map (OGM) extracted from a building information model (BIM model). However, most of these studies assume that the BIM model precisely represents the real world, which is rarely true. Discrepancies between the reference BIM model and the real world (Scan-BIM deviations) are not only due to furniture or clutter but also the usual as-planned and as-built deviations that exist with any model created in the design phase. These deviations affect the accuracy of AMCL drastically. This paper proposes an open-source method to generate appropriate Pose Graph-based maps from BIM models for robust 2D-LiDAR localization in changing and dynamic environments. First, 2D OGMs are automatically generated from complex BIM models. These OGMs only represent structural elements allowing indoor autonomous robot navigation. Then, an efficient technique converts these 2D OGMs into Pose Graph-based maps enabling more accurate robot pose tracking. Finally, we leverage the different map representations for accurate, robust localization with a combination of state-of-the-art algorithms. Moreover, we provide a quantitative comparison of various state-of-the-art localization algorithms in three simulated scenarios with varying levels of Scan-BIM deviations and dynamic agents. More precisely, we compare two Particle Filter (PF) algorithms: AMCL and General Monte Carlo Localization (GMCL); and two Graph-based Localization (GBL) methods: Google's Cartographer and SLAM Toolbox, solving the global localization and pose tracking problems. The numerous experiments demonstrate that the proposed method contributes to a robust localization with an as-designed BIM model or a sparse OGM in changing and dynamic environments, outperforming the conventional AMCL in accuracy and robustness.

cs.RO↗

arXiv2023

ECS -- an Interactive Tool for Data Quality Assurance

With the increasing capabilities of machine learning systems and their potential use in safety-critical systems, ensuring high-quality data is becoming increasingly important. In this paper we present a novel approach for the assurance of data quality. For this purpose, the mathematical basics are first discussed and the approach is presented using multiple examples. This results in the detection of data points with potentially harmful properties for the use in safety-critical systems.

cs.LG↗

arXiv2023

QI2 -- an Interactive Tool for Data Quality Assurance

The importance of high data quality is increasing with the growing impact and distribution of ML systems and big data. Also the planned AI Act from the European commission defines challenging legal requirements for data quality especially for the market introduction of safety relevant ML systems. In this paper we introduce a novel approach that supports the data quality assurance process of multiple data quality aspects. This approach enables the verification of quantitative data quality requirements. The concept and benefits are introduced and explained on small example data sets. How the method is applied is demonstrated on the well known MNIST data set based an handwritten digits.

cs.CY↗

arXiv2023

Windscreen Optical Quality for AI Algorithms: Refractive Power and MTF not Sufficient

Windscreen optical quality is an important aspect of any advanced driver assistance system, and also for future autonomous driving, as today at least some cameras of the sensor suite are situated behind the windscreen. Automotive mass production processes require measurement systems that characterize the optical quality of the windscreens in a meaningful way, which for modern perception stacks implies meaningful for artificial intelligence (AI) algorithms. The measured optical quality needs to be linked to the performance of these algorithms, such that performance limits - and thus production tolerance limits - can be defined. In this article we demonstrate that the main metric established in the industry - refractive power - is fundamentally not capable of capturing relevant optical properties of windscreens. Further, as the industry is moving towards the modulation transfer function (MTF) as an alternative, we mathematically show that this metric cannot be used on windscreens alone, but that the windscreen forms a novel optical system together with the optics of the camera system. Hence, the required goal of a qualification system that is installed at the windscreen supplier and independently measures the optical quality cannot be achieved using MTF. We propose a novel concept to determine the optical quality of windscreens and to use simulation to link this optical quality to the performance of AI algorithms, which can hopefully lead to novel inspection systems.

cs.CV↗

arXiv2023

Self-Supervised Multi-Object Tracking For Autonomous Driving From Consistency Across Timescales

Self-supervised multi-object trackers have tremendous potential as they enable learning from raw domain-specific data. However, their re-identification accuracy still falls short compared to their supervised counterparts. We hypothesize that this drawback results from formulating self-supervised objectives that are limited to single frames or frame pairs. Such formulations do not capture sufficient visual appearance variations to facilitate learning consistent re-identification features for autonomous driving when the frame rate is low or object dynamics are high. In this work, we propose a training objective that enables self-supervised learning of re-identification features from multiple sequential frames by enforcing consistent association scores across short and long timescales. We perform extensive evaluations demonstrating that re-identification features trained from longer sequences significantly reduce ID switches on standard autonomous driving datasets compared to existing self-supervised learning methods, which are limited to training on frame pairs. Using our proposed SubCo loss function, we set the new state-of-the-art among self-supervised methods and even perform on par with fully supervised learning methods.

cs.CV↗