Feifei Li

Papers

Total Citations

H-Index

About

Feifei Li is a leading researcher in robotics, computer vision, and multisensory object-centric learning, whose work bridges the gap between simulated environments and real-world applications. Her most impactful contribution is the development of the ObjectFolder 2.0 dataset (2022, 59 citations), a groundbreaking multisensory object dataset that virtualizes 100 objects with realistic visual, acoustic, and tactile properties. This work addresses a critical limitation in prior object modeling—unrealistic simulations—by enabling robust sim-to-real transfer for robotic manipulation and perception tasks. The dataset has become a foundational resource for researchers studying how robots can interact with objects using multiple senses, significantly advancing the field of embodied AI. Li’s more recent work, ARCap (2025, 7 citations), introduces an innovative method for collecting high-quality human demonstrations for robot learning using augmented reality feedback, showcasing her continued commitment to improving robot training efficiency and data quality. Her research has profound implications for developing robots that can seamlessly operate in human-centric environments, making her a pivotal figure in next-generation robotics. With her focus on multisensory integration and sim-to-real transfer, Li is shaping how machines perceive and interact with the physical world.