Tianfan Xue

Papers

Total Citations

H-Index

About

Tianfan Xue is a leading researcher at the forefront of embodied AI and 3D perception, with a focus on enabling robots to understand and interact with complex, real-world environments. His major contributions lie in developing holistic, multi-modal perception systems that bridge the gap between raw sensor data and actionable, language-grounded understanding. His highly influential work, "EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI" (2024), has already garnered 54 citations, establishing a foundational benchmark for embodied agents to explore, comprehend, and contextualize 3D scenes from first-person observations. This suite is critical for advancing robots that can follow human instructions in unstructured spaces. More recently, Xue has tackled a notoriously difficult problem in robotics with "FuseGrasp: Radar-Camera Fusion for Robotic Grasping of Transparent Objects" (2025). By pioneering a fusion of radar and camera data, his work overcomes the limitations of camera-only systems, enabling reliable robotic manipulation of transparent objects even in low-light conditions. Through these innovations, Xue is directly shaping the future of embodied intelligence, making robots more perceptive, robust, and capable in everyday human environments.