Papers
3
Total Citations
63
H-Index
3
About
Keyu Lu is a researcher whose work lies at the intersection of computer vision and robotics, with a particular focus on enabling machines to perceive and navigate their environments. Lu’s major contributions center on developing efficient deep learning architectures for vision-based tasks, most notably in object detection for robotic applications. Their 2017 paper on an efficient deep network for vision-based object detection, which has garnered 33 citations, has been influential in advancing real-time perception systems for robots. Prior to this, Lu made significant strides in field robot navigation with a 2015 study on vision sensor-based road detection, cited 24 times, which proposed a hierarchical method for robust environmental perception in challenging outdoor terrains. More recently, Lu has explored self-supervised learning for monocular depth estimation using quantized networks, a 2021 work that, while newer with 6 citations, demonstrates a forward-looking approach to reducing computational demands. Through these contributions, Lu has helped bridge the gap between theoretical computer vision and practical robotic systems, making autonomous navigation more reliable and efficient in real-world settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2Vision Sensor-Based Road Detection for Field Robot Navigation24 citations · 2015
- 3Self-supervised learning of monocular depth using quantized networks6 citations · 2021