Papers
3
Total Citations
72
H-Index
2
About
Xiaofeng Liu is a multidisciplinary researcher whose work spans computer vision, human-robot interaction, and speech processing, with a particular focus on developing intelligent systems that bridge perception and human communication. His most recognized contribution lies in the domain of autonomous driving, where his 2020 paper on importance-aware semantic segmentation introduced a novel discrete Wasserstein training framework to address limitations of conventional cross-entropy loss functions, achieving notable improvements in pixel-level scene understanding — a critical capability for self-driving vehicles and robotics. This work has garnered 54 citations, reflecting its meaningful impact on the computer vision community. Beyond autonomous systems, Liu has demonstrated a sustained commitment to socially impactful technology, developing an interactive motor-learning platform that enables children with autism spectrum disorder to engage in imitation-based training guided by robotic speech instructions, a contribution that bridges assistive technology and rehabilitation. His earlier work also explored the acoustic characteristics of stressed speech in human-robot interfaces, underscoring his broad interest in making machines more responsive to human variability. Taken together, Liu's research reflects a researcher dedicated to deploying intelligent systems in real-world, human-centered contexts.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3