Repository for Publications and Research Data
News from the ETH Library
Recently Added
Dynamic-0: Dynamic Neural Networks Applied to a Model-free 6D Detection of Unseen Objects Pipeline
Item type: Journal Article
Greco, Francesca Maria (2026)
Dynamic-0 (X0) addresses model-free 6D pose estimation for unseen objects—without Computer-Aided Design (CAD) models—while remaining efficient across varied hardware. It uses a dual-agent design: the first agent achieves coarse pose alignment via DINOv3 visual descriptors, Facebook Artificial Intelligence Similarity Search (FAISS)-based retrieval, and Efficient Perspective-n-Point (EPnP) with Super Random Sample Consensus (SuperRANSAC); the second selectively refines this estimate using differentiable 3D Gaussian Splatting (3DGS) rendering, improving geometric precision only where needed to avoid uniform computational overhead. To adapt inference dynamically, X0 integrates multi-scale routing, sparse feature reactivation, and a novelty-aware semantic gating mechanism based on generative density scoring over DINOv3 and MesaNet embeddings. A MesaNet temporal block further stabilizes pose sequences by treating temporal consistency as a locally optimal test-time training problem, solved via conjugate-gradient updates. Evaluated under the Benchmark for 6D Object Pose Estimation (BOP) Challenge 2025 model-free protocol on BOP-H3 datasets (HOPEv2, HANDAL, HOT3D), Dynamic-0 achieves an Average Precision of 0.473 on HOPEv2, with adaptive throughput of 2.27–15.4 Frames Per Second (FPS) and a 73% reduction in temporal jitter (average processing time: 40.98 s/image). Additional results are reported on HANDAL and HOT3D leaderboards, scoring as the first one in terms of accurate precision and average time on the leaderboards of all three datasets of the BOP Challenge 2025. Code is publicly released at the GitHub repository; the official leaderboards are available at the links indicated in the Data Availability section of this paper. These findings suggest that combining self-supervised representations, adaptive computation, generative novelty detection, and temporal optimization offers a practical route toward scalable, hardware-aware 6D pose estimation in real-world, model-free settings.
Long-term performance monitoring of a secondary electrospray ionisation high-resolution mass spectrometry platform using a vapor generator for routine quality control
Item type: Journal Article
Vadakkechira, Albin; Zenobi, Renato; Pohl, Daniel; et al. (2026)
Secondary electrospray ionisation high-resolution mass spectrometry (SESI-HRMS) is a powerful, non-invasive tool for real-time breath analysis. However, instrumental drift, capillary degradation and other technical variations compromise clinical applications, especially for long-term studies, which require robust quality control (QC) methodologies. In this work, we present the use of an evaporation chamber-based reference gas system to monitor SESI-MS system performance over almost two years. Targeted QC measurements with deuterated butyric acid were recorded three times on each measurement day. Analytical stability was systematically quantified across both positive and negative ion modes. In a standalone continuous characterization, the reference gas generator delivered a stable signal over several hours. A Linear Mixed-Effects Model (LMM) was constructed to evaluate the influence of time-of-day on instrument performance. This revealed significant effects: positive mode signal intensity progressively increased throughout the day, whereas for negative mode, an initial drop in signal intensity and subsequent stabilisation was observed. Furthermore, the longitudinal QC framework successfully tracked capillary performance over time, providing objective performance metrics to retrospectively evaluate the impact of targeted hardware maintenance on system performance. The implementation of a structured QC protocol provides a valid methodology for long-term SESI-HRMS performance monitoring, ensuring data integrity and supporting the translation of SESI-HRMS into large-scale clinical applications.
GBF-Indicator: A.4 Proportion of populations within species with an effective population size greater than 500
Item type: Dataset
Tschan, Jasmine Noëlle; Fischer, Martin C.; Priklopil, Tadeas; et al. (2026)
Tschan, Jasmine Noëlle; Fischer, Martin C.; Priklopil, Tadeas; et al. (2026)
Sustainability Assessment of Low-Carbon Fuels Towards Net Zero Emissions
Item type: Doctoral Thesis
Liu, Zipeng (2026)