About me

I’m Yixuan Hu, an M.Sc. graduate in Robotics, Cognition and Intelligence from the Technical University of Munich. I completed my bachelor’s degree in automotive engineering at Tongji University, Shanghai. My research interests lie in 3D computer vision, multimodal learning and generative models. For my master thesis, under the supervison of PD Dr. Slobodan Ilic and Dekai Zhu I developed a diffusion-based model for LiDAR data generation and introduced unified metrics to evaluate geometric and semantic quality of generated LiDAR scenes, enabling controllable LiDAR scene synthesis. This work was accepted to NeurIPS 2025. Besides, in the project ControlEvents, I achiveved controllable event-camera data generation, where I had the opportunity to work under the co-supervision of Prof. Daniel Cremers and Prof. Gerard Pons-Moll.

Publications

• SPIRAL: Semantic-Aware Progressive LiDAR Scene Generation (NeurIPS 2025) [Project Page] [Paper]
• ControlEvents: Controllable Synthesis of Event Camera Datawith Foundational Prior from Image Diffusion Models (Preprint) [Project Page] [Paper]