Papers
23
Total Citations
235
H-Index
10
About
Liyuan Li is a computer vision researcher whose work sits at the intersection of robotics perception, human-robot interaction, and intelligent visual systems. Over more than a decade of sustained scholarship, Li has made significant contributions to enabling mobile service and social robots to perceive, understand, and interact with people in real-world public environments. Li's most influential work focuses on multiperson detection and tracking, with a landmark 2012 system that introduced a maximum likelihood-based fusion algorithm combining multiple vision models to robustly identify and follow people in dynamic settings (31 citations). Complementing this, Li has advanced addressee selection — helping robots determine whom to engage during multi-party interactions (20 citations) — and developed HOG-based multi-stage frameworks for object detection and pose recognition at practical computational costs (19 citations). Early foundational contributions include stereo-based human detection (2005) and vision-based lift-button recognition to support autonomous robot navigation across building floors. More recently, Li has extended this expertise into 6D pose estimation using RGB-D fusion and adaptive multiview active perception, reflecting an ongoing engagement with modern deep learning approaches. With a body of work spanning over 150 cumulative citations, Li's research meaningfully shapes how robots see and socially navigate the human world.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5Lift-button detection and recognition for service robot in buildings15 citations · 2009
- 6Stereo-based human detection for mobile service robots14 citations · 2005
- 7
- 8
- 9Towards Efficient Multiview Object Detection with Adaptive Action Prediction10 citations · 2021
- 106D Pose Estimation with Correlation Fusion10 citations · 2021