Piano
I’ve played piano for over 14 years and competed nationally.
I’m an ELLIS PhD student at the Technical University of Munich, working with Prof. Nassir Navab at the Chair of Computer-Aided Medical Procedures (CAMP).
My research focuses on understanding people and scenes in 3D from multiple cameras. I work on multi-camera perception and generative modeling, with applications in the operating room.
Away from research, you’ll usually find me at the piano or in the pool. I also occasionally work on call as a trained paramedic.
Interested in a thesis project or collaboration in these areas? Feel free to reach out at tony.wang@tum.de.
A selection of my work as first or joint first author. A complete publication list is available on Google Scholar.
Estimates 3D human poses from multiple cameras without training a model. It matches detections jointly across views using a hypergraph, making the associations more robust to occlusion and noisy detections.
@inproceedings{wang2026compose,
title={COMPOSE: Hypergraph Cover Optimization for Multi-view 3D Human Pose Estimation},
author={Wang, Tony Danjun and Birdal, Tolga and Navab, Nassir and Bastian, Lennart},
booktitle={Advances in Neural Information Processing Systems (NeurIPS)},
year={2026}
}Represents the operating room using higher-order relationships rather than only pairwise links. It brings 3D geometry, audio, and robot kinematics into one representation while preserving their distinct structures.
@inproceedings{wang2026topoor,
title={TopoOR: A Unified Topological Scene Representation for the Operating Room},
author={Wang, Tony Danjun and Kim, Ka Young and Birdal, Tolga and Navab, Nassir and Bastian, Lennart},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI)},
year={2026},
organization={Springer}
}Learns to estimate 3D human poses from multiple cameras without 3D supervision. It treats matching people across camera views as a diffusion process and works across different camera setups, including scenes with heavy occlusion.
@inproceedings{wang2026dispose,
title={DisPOSE: Projected Polystochastic Diffusion for Self-Supervised Multi-View 3D Human Pose Estimation},
author={Wang, Tony Danjun and Birdal, Tolga and Navab, Nassir and Bastian, Lennart},
booktitle={International Conference on Machine Learning (ICML)},
year={2026}
}Identifies individual staff in the operating room from their 3D shape and motion instead of their appearance, which barely changes under standardized scrubs. The goal is to stop treating people as interchangeable roles and start modeling who they are and how they work together.
@article{wang2025beyond,
title={Beyond Role-Based Surgical Domain Modeling: Generalizable Re-Identification in the Operating Room},
author={Wang, Tony Danjun and Bastian, Lennart and Czempiel, Tobias and Heiliger, Christian and Navab, Nassir},
journal={Medical Image Analysis},
volume={105},
pages={103687},
year={2025},
publisher={Elsevier},
doi={10.1016/j.media.2025.103687}
}Anonymizes faces in operating room recordings by combining multiple camera views into a 3D scene and rendering replacement faces consistently across views, helping locate faces that 2D methods miss under occlusion.
@article{bastian2023disguisor,
title={DisguisOR: Holistic Face Anonymization in the Operating Room},
author={Bastian, Lennart and Wang, Tony Danjun and Czempiel, Tobias and Busam, Benjamin and Navab, Nassir},
journal={International Journal of Computer Assisted Radiology and Surgery},
note={Presented at IPCAI 2023},
year={2023},
publisher={Springer}
}@inproceedings{wang2025trackor,
title={TrackOR: Towards Personalized Intelligent Operating Rooms Through Robust Tracking},
author={Wang, Tony Danjun and Heiliger, Christian and Navab, Nassir and Bastian, Lennart},
booktitle={International Workshop on Collaborative Intelligence and Autonomy in Image-Guided Surgery},
pages={53--63},
year={2025},
organization={Springer}
}@article{wang2025mitigating,
title={Mitigating Biases in Surgical Operating Rooms with Geometry},
author={Wang, Tony Danjun and Czempiel, Tobias and Navab, Nassir and Bastian, Lennart},
journal={arXiv preprint arXiv:2508.08028},
year={2025}
}I’ve played piano for over 14 years and competed nationally.
I’m trained and certified as a paramedic in Germany and occasionally work on call.
I used to coach swimming. These days, I’m still in the pool most days.
I spent several years tutoring computer science and maths fundamentals.
I prefer seeing films on the big screen. Going to the cinema is one of my favorite ways to spend an evening.