Lili Chen

Shanghai Institute of Microsystem and Information Technology, Technical University of Munich

Papers

Total Citations

H-Index

About

Lili Chen is a robotics researcher whose work sits at the intersection of computer vision, human-robot interaction, and robot learning. Her key research areas include 3D hand pose estimation, multi-person tracking, and skill acquisition from unstructured data. Chen’s major contributions include developing a generic Topology-aware Transformer model for 3D hand pose and mesh estimation, which addresses the critical challenge of severe self-occlusion and high self-similarity in human-robot interaction contexts. She also proposed a global optimization approach for multiple people tracking, improving robustness in video surveillance and HRI applications. Her work on PlayFusion introduces diffusion-based skill acquisition from language-annotated play data, enabling robots to learn from unstructured, uncurated behavior data—a paradigm shift from traditional structured learning. Additionally, Chen’s research on leveraging affordances from human videos provides a versatile representation for robotics, bridging the gap between static datasets and real-world robot interaction. With papers published in 2023-2024, her work is gaining traction, accumulating citations that reflect growing interest in her innovative approaches to making robots more capable of understanding and interacting with humans in natural, unstructured environments.