Xiaopeng Wei
Papers
13
Total Citations
231
H-Index
6
About
Xiaopeng Wei is a versatile researcher whose work spans computer vision, robotics, and human-robot interaction, with particular expertise in visual perception and autonomous navigation. His most influential contribution — the 2020 paper "Don't Hit Me! Glass Detection in Real-World Scenes" — tackled a deceptively difficult problem: enabling computer vision systems to reliably detect transparent glass surfaces, which had been largely overlooked despite posing serious safety risks to autonomous robots and vehicles. Garnering 131 citations, this work established a foundational benchmark in the field, and Wei extended it further in a 2022 follow-up exploring large-field contextual feature learning for improved glass detection. Earlier in his career, Wei made significant contributions to robotic path planning, developing hybrid genetic algorithm and A* algorithm approaches for efficient mobile robot navigation, work that collectively attracted dozens of citations. His research portfolio also encompasses space manipulator trajectory planning, human motion prediction using recurrent neural networks, anomaly detection for safe human-robot interaction, and surgical robotics with visual question answering. Across these diverse domains, Wei consistently bridges perception, planning, and safety — addressing real-world challenges that matter deeply for the next generation of intelligent robotic systems.
Research Focus
Key Achievements
Top Papers
- 1Don’t Hit Me! Glass Detection in Real-World Scenes131 citations · 2020
- 2GA-based Global Path Planning for Mobile Robot Employing A* Algorithm25 citations · 2012
- 3Large-Field Contextual Feature Learning for Glass Detection20 citations · 2022
- 4
- 5Fusion learning-based recurrent neural network for human motion prediction11 citations · 2022
- 6Optimal Path Planning for Minimizing Base Disturbance of Space Robot7 citations · 2016
- 7
- 8Asymmetric Anomaly Detection for Human-Robot Interaction3 citations · 2021
- 9
- 10A Novel Gaze-Point-Driven HRI Framework for Single-Person2 citations · 2021