Xiaohan Yan
Papers
1
Total Citations
1
H-Index
1
About
Xiaohan Yan is a rising researcher at the forefront of 3D computer vision and robotics, whose work tackles the fundamental challenge of enabling machines to recognize and segment objects in three-dimensional space. Their most notable contribution, "RE0: Recognize Everything with 3D Zero-Shot Instance Segmentation," introduces a groundbreaking approach that leverages vision foundation models (VFMs) from 2D to overcome the persistent scarcity of high-quality 3D training data. This work, published in 2025, pioneers a zero-shot paradigm for 3D instance segmentation, allowing robots to identify novel objects without prior 3D-specific training—a critical step toward general-purpose robotic perception. Yan's research directly addresses the infeasibility of training large-scale 3D models from scratch, instead repurposing the rich representations of 2D VFMs to bridge the domain gap. While still early in their career, Yan's contributions are already shaping the future of embodied AI, offering a scalable path to "recognize everything" in the real world. Their work holds immense promise for applications in autonomous navigation, manipulation, and scene understanding, positioning Yan as a key innovator in the intersection of foundation models and 3D vision.
Research Focus
Key Achievements
Top Papers
- 1RE0: Recognize Everything with 3D Zero-Shot Instance Segmentation1 citations · 2025