| Name: | David Tschirschwitz, M.Sc. | |
| Raum: | Bauhausstraße 11, Raum 009 | |
| Telefon: | - | |
| E-Mail: | david.tschirschwitz[at]uni-weimar.de | |
| Sprechstunde: | Auf Anfrage | |
| Lehre: | Deep Learning for Computer Vision (SS22, SS23) ReTMeD - Replication and Transformation of Medical Object Detection (WS24/25) BUWLense: AI-Powered Image-to-Image Search (SS24) RODSL - Learning Robust Object Detection with Soft-Labels from Multiple Annotators (WS22/23) | |
| Forschung: | Human Label Variation, Object Detection, Instance Segmentation, Image Retrieval, Document Analysis and Recognition | |
Aktuelle Forschungsprojekte
- VarInsightAI - Sichtbarmachung und Bewältigung realer Annotationsvariation in KI-gestützter Bildanalyse
Publications
- D. Tschirschwitz & V. Rodehorst: Label Convergence: Defining an Upper Performance Bound in Object Recognition through Contradictory Annotations, In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025. (Accepted manuscript) [arxiv]
- D. Tschirschwitz & V. Rodehorst: CISOL: An Open and Extensible Dataset for Table Structure Recognition in the Construction Industry, In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025. (Accepted manuscript) [Dataset][Leaderboard]
- M. Florek, D. Tschirschwitz, B. Barz & V. Rodehorst: Efficient and Discriminative Image Feature Extraction for Universal Image Retrieval, In: Proceedings of the German Conference on Pattern Recognition (GCPR), 2024. (In print) [Repository]
- D. Tschirschwitz, C. Benz, M. Florek, H. Norderhus, B. Stein & V. Rodehorst: Drawing the Same Bounding Box Twice? Coping Noisy Annotations in Object Detection with Repeated Labels, In: Proceedings of the German Conference on Pattern Recognition (GCPR), 2023. [arxiv][doi]
- D. Tschirschwitz, F. Klemstein, H. Schmidgen & V. Rodehorst: Drawing the Line: A Dual Evaluation Approach for Shaping Ground Truth in Image Retrieval Using Rich Visual Embeddings of Historical Images, 7th International ICDAR-Workshop on Historical Document Imaging and Processing (HIP), 2023. [doi][DHMTIC][DHREAAL]
Supervised Theses
- Gekeler, Ella Maria: Construction Industry Steel Ordering List (CISOL) Dataset: A Data- Driven Approach to Table Structure Recognition [Published at WACV25]
- Florek, Morris: Efficient and Discriminative Image Feature Extraction for Multi-Domain Image Retrieval [Published at the GCPR2024 YRF]
- Yarakaraju, Raghu Varma: Finding Image-to-Image Similarities in a Historic Humanities Corpus
- Steinhaußen, Julian: Deep Learning for Image Registration