I am an ML researcher specializing in hyperspectral image analysis and anomaly detection, currently working at the intersection of fundamental research and real-world application at Fraunhofer IGD and the University of Rostock.
My work centers on a problem that matters for industrial inspection, medical imaging and beyond: building models that detect anomalies reliably across different materials, sensors and acquisition environments. I work with Specim FX10/FX17 and Cubert VNIR hyperspectral cameras and develop deep learning methods for spectral-spatial representation learning, band selection and cross-domain generalization.
- 🔬 Currently developing HSI pipelines for physiological monitoring and anomaly detection
- 🧠 Exploring self-supervised representation learning and meta-learning for cross-domain HSI generalization
- 📄 Published in Journal of Imaging (2026) — hematoma age estimation from hyperspectral images using CNNs
- 🎓 M.Sc. Information Technology, University of Rostock (expected August 2026)
|
Spatial-spectral CNN for hematoma age estimation from HSI medical images Band selection: 204 channels reduced to task-relevant subset +30% accuracy over baseline Published · Journal of Imaging (2026) |
Investigating domain shift across subjects and acquisition conditions Exploring self-supervised representation learning Meta-learning for robust cross-site anomaly models Ongoing · Fraunhofer IGD & University of Rostock |
ML pipeline for thigh muscle force prediction from EIT sensor data Custom 16-electrode wearable sensor belt Physics-informed signal processing and deep learning M.Sc. Thesis · University of Rostock |
AI, ML & Deep Learning
Hyperspectral & Computer Vision
MLOps & Cloud
Languages & Tools