谢靖超
Ph.D. Student · Self-Supervised 3D Vision
Visual Inference Lab, DFG Cluster of Excellence "Reasonable AI", TU Darmstadt
I am a Ph.D. student in the DFG Cluster of Excellence "Reasonable AI", TU Darmstadt, working on self-supervised 3D understanding.
Previously, I worked with the Computer Vision Group and DeepScenario at the Technical University of Munich on unsupervised monocular visual odometry, and with Bosch Research on unsupervised online mapping for autonomous driving.
I received my M.Sc. in Robotics, Cognition, Intelligence from the Technical University of Munich and my B.Sc. in Vehicle Engineering from Central South University in China.
International Journal of Computer Vision (IJCV), 2026, to appear.
Journal extension of CoProU-VO, studying projected uncertainty for robust self-supervised visual odometry from two-frame to multi-frame settings.
Paper link coming soon
Proceedings of the DAGM German Conference on Pattern Recognition (GCPR), 2025.
An end-to-end unsupervised monocular visual odometry method that uses projected uncertainty to improve robustness under dynamic objects, occlusions, and non-Lambertian regions.
arXiv / Project Page / Award
Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2025.
A framework for learning online mapping without HD map supervision, developed during my research internship at Bosch Research.