Lectures 2026/27
- Deep Learning – Master in Computing for Data Science (3rd semester)
- Computer Vision – Master in Computing for Data Science (4th semester)
- AI and Machine Learning – Bachelor in Electronic and Information Engineering (5th semester)
- Introduction to Computer Vision – Bachelor in Electronic and Information Engineering (6th semester)
MSc and BSc Theses
Current thesis topics
- Contrastive learning for video action anticipation
- Temporal action detection in video streams
- Computer vision for workplace safety
- Computer vision for team sports
Completed theses:
- Human Temporal Action Detection in Video Streams
- Evaluation of Action Recognition Architectures for a Production Environment by the example of Krone GmbH & Co
- Fine-grained Action Recognition in Egocentric Videos (with 1st place at HoloAssist Challenge, CVPR 2024)
- Foundation Models applied to Depth Estimation for SLAM problem (with internship at Covision Lab)
- Automated Evaluation and Classification of Fat Bloom on Chocolate Surfaces using Image Processing Algorithms and Machine Learning Techniques (with internship at Loacker)
- Retrieval-Augmented Generation Chatbot: A Proof-of-Concept Implementation for Customer Service at Durst Group AG (with internship at Durst)
- Improving Semantic Segmentation Models for Vision-Based Anomaly Detection with Synthetic Data from Diffusion Models (with workshop paper at ICLR 2025)
- Human Action Recognition from Egocentric Video (with 2nd place at MECCANO Challenge, ICIAP 2023)
- Utilizing Multi-modal Data for Human Action Recognition Tasks
- Performance Disparities and Improvement Strategies in Industrial Anomaly Detection: A Comparative Case Study using MVTech and Real-World Data
- Real-Time Human 3D Pose Tracking by Fusing Egocentric and Exocentric Video Streams