Songlin Wei
Papers
3
Total Citations
18
H-Index
2
About
Songlin Wei is a rising researcher at the intersection of computer vision, robotics, and semantic perception. Their work centers on two ambitious frontiers: enabling robots to understand and manipulate deformable objects, and advancing semantic simultaneous localization and mapping (SLAM) for autonomous systems. Wei’s foundational contribution to SLAM appears in their most-cited paper, “Object Clustering With Dirichlet Process Mixture Model for Data Association in Monocular SLAM” (12 citations), which tackles the critical data association problem—assigning persistent identities to objects across frames—using probabilistic clustering to make monocular semantic SLAM more robust and practical. In robotic manipulation, Wei has pioneered zero-shot and generalizable approaches for deformable objects. Their work “Make a Donut” (4 citations) introduces hierarchical Earth Mover’s Distance planning for tool-based manipulation without task-specific training, while “RoboHanger” (2 citations) addresses the underexplored challenge of inserting hangers into diverse garments laid flat—a deceptively complex task requiring fine-grained perception and control. These contributions demonstrate Wei’s commitment to pushing beyond rigid-object assumptions, tackling real-world variability with elegant, learning-free strategies. For students and researchers, Wei’s trajectory offers a compelling model of how to bridge geometric reasoning and robotic dexterity.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3