Ke-Jyun Wang
Papers
2
Total Citations
16
H-Index
2
About
Ke-Jyun Wang is a researcher working at the intersection of computer vision, natural language processing, and robotics, with a particular focus on visual grounding and embodied AI. His most notable contribution is the development of **OCID-Ref**, a pioneering 3D robotic dataset that pairs cluttered scene imagery with embodied language to advance research in visual grounding for occluded objects. This work, published at the prestigious NAACL 2021 conference and accumulating over 14 citations, addresses a critical gap in the field: the lack of realistic, working-environment datasets — such as offices and warehouses — where robots must identify and interact with partially hidden objects using natural language instructions. Wang's research tackles a genuinely challenging problem: enabling robots to understand referring expressions in cluttered, real-world environments where occlusion significantly complicates object recognition. By constructing a dataset grounded in 3D spatial understanding and natural language, he has provided the broader research community with a valuable benchmark for evaluating and improving human-robot interaction systems. His work sits at a compelling crossroads of embodied AI and language grounding, areas increasingly vital as autonomous systems become more integrated into everyday working environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2