arXiv subjects

Dominik Schneider

Publications and source records attributed to Dominik Schneider.

7 recordsLinked to original sources

arXiv2026

SAW: Toward a Surgical Action World Model via Controllable and Scalable Video Generation

A surgical world model capable of generating realistic surgical action videos with precise control over tool-tissue interactions can address fundamental challenges in surgical AI and simulation -- from data scarcity and rare event synthesis to bridging the sim-to-real gap for surgical automation. However, current video generation methods, the very core of such surgical world models, require expensive annotations or complex structured intermediates as conditioning signals at inference, limiting their scalability. Other approaches exhibit limited temporal consistency across complex laparoscopic scenes and do not possess sufficient realism. We propose Surgical Action World (SAW) -- a step toward surgical action world modeling through video diffusion conditioned on four lightweight signals: language prompts encoding tool-action context, a reference surgical scene, tissue affordance mask, and 2D tool-tip trajectories. We design a conditional video diffusion approach that reformulates video-to-video diffusion into trajectory-conditioned surgical action synthesis. The backbone diffusion model is fine-tuned on a custom-curated dataset of 12,044 laparoscopic clips with lightweight spatiotemporal conditioning signals, leveraging a depth consistency loss to enforce geometric plausibility without requiring depth at inference. SAW achieves state-of-the-art temporal consistency (CD-FVD: 199.19 vs. 546.82) and strong visual quality on held-out test data. Furthermore, we demonstrate its downstream utility for (a) surgical AI, where augmenting rare actions with SAW-generated videos improves action recognition (clipping F1-score: 20.93% to 43.14%; cutting: 0.00% to 8.33%) on real test data, and (b) surgical simulation, where rendering tool-tissue interaction videos from simulator-derived trajectory points toward a visually faithful simulation engine.

cs.CV↗

arXiv2026

Towards Controllable Video Synthesis of Routine and Rare OR Events

Purpose: Curating large-scale datasets of operating room (OR) workflow, encompassing rare, safety-critical, or atypical events, remains operationally and ethically challenging. This data bottleneck complicates the development of ambient intelligence for detecting, understanding, and mitigating rare or safety-critical events in the OR. Methods: This work presents an OR video diffusion framework that enables controlled synthesis of rare and safety-critical events. The framework integrates a geometric abstraction module, a conditioning module, and a fine-tuned diffusion model to first transform OR scenes into abstract geometric representations, then condition the synthesis process, and finally generate realistic OR event videos. Using this framework, we also curate a synthetic dataset to train and validate AI models for detecting near-misses of sterile-field violations. Results: In synthesizing routine OR events, our method outperforms off-the-shelf video diffusion baselines, achieving lower FVD/LPIPS and higher SSIM/PSNR in both in- and out-of-domain datasets. Through qualitative results, we illustrate its ability for controlled video synthesis of counterfactual events. An AI model trained and validated on the generated synthetic data achieved a RECALL of 70.13% in detecting near safety-critical events. Finally, we conduct an ablation study to quantify performance gains from key design choices. Conclusion: Our solution enables controlled synthesis of routine and rare OR events from abstract geometric representations. Beyond demonstrating its capability to generate rare and safety-critical scenarios, we show its potential to support the development of ambient intelligence models.

cs.CV↗

arXiv2025

Quasi-optimal error estimates for the approximation of stable stationary states of the elastic energy of inextensible curves

We establish local existence and a quasi-optimal error estimate for piecewise cubic minimizers to the bending energy under a discretized inextensibility constraint. In previous research a discretization is used where the inextensibility constraint is only enforced at the nodes of the discretization. We show why this discretization leads to suboptimal convergence rates and we improve on it by also enforcing the constraint in the midpoints of each subinterval. We then use the inverse function theorem to prove existence and an error estimate for stationary states of the bending energy that yields quasi-optimal convergence. We use numerical simulations to verify the theoretical results experimentally.

math.NA↗

arXiv2025

Quasi-optimal error estimate for the approximation of the elastic flow of inextensible curves

A space-discretization for the elastic flow of inextensible curves is devised and quasi-optimal convergence of the corresponding semi-discrete problem is proved for a suitable discretization of the nonlinear inextensibility constraint. Further a fully discrete time-stepping scheme that incorporates this constraint is proposed and unconditional stability and convergence of the discrete scheme are proved. Finally some numerical simulations are used to verify the obtained results experimentally.

math.NA↗

arXiv2022

Orchestrating Tool Chains for Model-based Systems Engineering with RCE

When using multiple software tools to analyze, visualize, or optimize models in MBSE, it is often tedious and error-prone to manually coordinate the execution of these tools and to retain their respective input and output data for later analysis. Since such tools often require expertise in their usage as well as diverse run-time environments, it is not straightforward to orchestrate their execution via off-the-shelf software tools. We present RCE, an application developed at the German Aerospace Center that supports engineers in developing and orchestrating the execution of complex tool chains. This application is used in numerous research and development projects in diverse domains and enables and simplifies the creation, analysis, and optimization of models.

cs.SE↗

arXiv2022

Automated and manual testing as part of the research software development process of RCE

Research software is often developed by individual researchers or small teams in parallel to their research work. The more people and research projects rely on the software in question, the more important it is that software updates implement new features correctly and do not introduce regressions. Thus, developers of research software must balance their limited resources between implementing new features and thoroughly testing any code changes. We present the processes we use for developing the distributed integration framework RCE at DLR. These processes aim to strike a balance between automation and manual testing, reducing the testing overhead while addressing issues as early as possible. We furthermore briefly describe how these testing processes integrate with the surrounding processes for development and releasing.

cs.SE↗

arXiv2021

Towards Automated Semantic Grouping in Workflows for Multi-Disciplinary Analysis

When designing multidisciplinary tool workflows in visual development environments, researchers and engineers often combine simulation tools which serve a functional purpose and helper tools that merely ensure technical compatibility by, e.g., converting between file formats. If the development environment does not offer native support for such groups of tools, maintainability of the developed workflow quickly deteriorates with an increase in complexity. We present an approach towards automatically identifying such groups of closely related tools in multidisciplinary workflows implemented in RCE by transforming the workflow into a graph and applying graph clustering algorithms to it. Further, we implement this approach and evaluate multiple clustering algorithms. Our results strongly indicate that this approach can yield groups of closely related tools in RCE workflows, but also that solutions to this problem will have to be tailor-made to each specific style of workflow design.

cs.SE↗