3D reconstruction of dynamic scenes.

Publications

1. 2026 Journal article

   

   OmniPrior: A Multi-Prior-Guided Omnidirectional Representation of Dynamic Scenes in Overlapping Ultra-Wide Multi-Fisheye Videos

   Simin Kou, Fang-Lue Zhang, Jakob Nazarenus, Reinhard Koch, Can Wang, Neil A. Dodgson

   IEEE Transactions on Visualization and Computer Graphics

   Omnidirectional capture of dynamic scenes facilitates the creation of immersive virtual reality assets and holistic scene understanding. Outward-facing multi-fisheye camera rigs offer an efficient solution for full-scene coverage, using fewer lenses than conventional pinhole arrays while enabling all-directional observation of…

   Full text → dynamic scenes · omnidirectional · priors
2. 2025 Conference paper

   

   2D Sheet Temperature Estimation for Multi-Stage Press Hardening in a Progressive Die

   Henry Baumann, Jakob Nazarenus, Juri Martschin, A. Erman Tekkaya, Thomas Meurer

   IFAC-PapersOnLine

   During a multi-stage press hardening process, where a metal sheet undergoes rapid austenitization, tempering, and forming, the spatial-temporal temperature development is a driving factor and its control enables it to reach desired product properties. A two-dimensional model extension is proposed, which enables a precise and…

   Full text → estimation · manufacturing · thermal
3. 2025 Preprint

   

   Gaussians on Fire: High-Frequency Reconstruction of Flames

   Jakob Nazarenus, Dominik Michels, Wojtek Pałubicki, Simin Kou, Fang-Lue Zhang, Sören Pirk, Reinhard Koch

   arXiv

   We propose a method to reconstruct dynamic fire in 3D from a limited set of camera views with a Gaussian-based spatiotemporal representation. Capturing and reconstructing fire and its dynamics is highly challenging due to its volatile nature, transparent quality, and multitude of high-frequency features. Despite these challenges, we…

   Full text → Gaussian splatting · dynamic reconstruction · fire
4. 2025 Journal article

   Occlusion Avoidance for Robotic Manipulators Using Rigid Gaussian Splatting

   Jakob Nazarenus, Simon Reichhuber, Reinhard Koch, Sven Tomforde, Simin Kou, Fang-Lue Zhang

   SN Computer Science

   Monitoring robotic manipulators is essential in highly automated environments, where optical cameras can provide precise and dense information while suffering from line-of-sight occlusions. This paper introduces a lightweight, learning-based solution to address these occlusions in robot-camera systems. We employ a small MLP…

   Full text → Code → robotics · Gaussian splatting · manipulation
5. 2025 Journal article

   

   OmniPlane: A Recolorable Representation for Dynamic Scenes in Omnidirectional Videos

   Simin Kou, Fang-Lue Zhang, Jakob Nazarenus, Reinhard Koch, Neil A. Dodgson

   IEEE Transactions on Visualization and Computer Graphics

   Consumer-level omnidirectional video offers an economically viable means to create virtual reality (VR) assets, enabling users to explore and interact within a fully immersive visual environment. However, editing such videos, particularly those with 360° views and dynamic objects, poses significant challenges. Existing approaches to…

   Full text → dynamic scenes · omnidirectional · representation
6. 2024 Journal article

   Arbitrary Optics for Gaussian Splatting Using Space Warping

   Jakob Nazarenus, Simin Kou, Fang-Lue Zhang, Reinhard Koch

   Journal of Imaging

   Due to recent advances in 3D reconstruction from RGB images, it is now possible to create photorealistic representations of real-world scenes that only require minutes to be reconstructed and can be rendered in real time. In particular, 3D Gaussian splatting shows promising results, outperforming preceding reconstruction methods while…

   Full text → Code → Gaussian splatting · camera models · optics
7. 2024 Journal article

   

   Hydrogel System with Independent Tailoring of Mechanics, CT, and US Contrasts for Affordable Medical Phantoms

   Haoyi Qiu, Jakob Nazarenus, Bernhard Egeler, Tom Thode, Firdaws Osman, Daniar Osmonov, Jörg Bahr, Sören Kaps, Frank-André Siebert, Reinhard Koch, Ulf Lützen, Rainer Adelung, Leonard Siebert

   ACS Materials Letters

   Medical phantoms mimic aspects of procedures like computed tomography (CT), ultrasound (US) imaging, and surgical practices. However, the materials for current commercial phantoms are expensive and the fabrication with these is complex and lacks versatility. Therefore, existing material solutions are not suitable for creating…

   Full text → medical imaging · phantoms · materials
8. 2024 Book chapter

   

   Label Smarter, Not Harder: CleverLabel for Faster Annotation of Ambiguous Image Classification with Higher Quality

   Lars Schmarje, Vasco Grossmann, Tim Michels, Jakob Nazarenus, Monty Santarossa, Claudius Zelenka, Reinhard Koch

   Lecture Notes in Computer Science (GCPR)

   High-quality data is crucial for the success of machine learning, but labeling large datasets is often a time-consuming and costly process. While semi-supervised learning can help mitigate the need for labeled data, label quality remains an open issue due to ambiguity and disagreement among annotators. Thus, we use proposal-guided…

   Full text → Code → annotation · classification · data quality
9. 2024 Conference paper

   Best Paper Nominee

   Learning Occlusions in Robotic Systems: How to Prevent Robots from Hiding Themselves

   Jakob Nazarenus, Simon Reichhuber, Manuel Amersdorfer, Lukas Elsner, Reinhard Koch, Sven Tomforde, Hossam S. Abbas

   ICAART

   In many applications, robotic systems are monitored via camera systems. This helps with monitoring automated production processes, anomaly detection, and the refinement of the estimated robot’s pose via optical tracking systems. While providing high precision and flexibility, the main limitation of such systems is their line-of-sight constraint. In this paper, we propose a lightweight solution for automatically learning this occluded space to provide continuously observable robot trajectories. This is achieved by an initial autonomous calibration procedure and subsequent training of a simple neural network. During operation, this network provides a prediction of the visibility status with a balanced accuracy of 90% as well as a gradient that leads the robot to a more well-observed area. The prediction and gradient computations run with sub-ms latency and allow for modular integration into existing dynamic trajectory-planning algorithms to ensure high visibility of the desired target.

   Full text → Code → robotics · occlusion · learning
10. 2024 Journal article

    

    When Two Eyes Don’t Suffice - Learning Difficult Hyperfluorescence Segmentations in Retinal Fundus Autofluorescence Images via Ensemble Learning

    Monty Santarossa, Tebbo Tassilo Beyer, A. Scharf, Ayse Tatli, Claus von der Burchard, Jakob Nazarenus, Johann Roider, Reinhard Koch

    Journal of Imaging

    Hyperfluorescence (HF) and reduced autofluorescence (RA) are important biomarkers in fundus autofluorescence images (FAF) for the assessment of health of the retinal pigment epithelium (RPE), an important indicator of disease progression in geographic atrophy (GA) or central serous chorioretinopathy (CSCR). Autofluorescence images…

    Full text → medical imaging · segmentation · ensemble