Papers
2
Total Citations
54
H-Index
2
About
Yidan Feng is a leading researcher in industrial robotics, specializing in sim-to-real transfer learning for object recognition and localization. Their work addresses a critical bottleneck in automated bin picking—the gap between simulated training environments and real-world industrial applications. Feng's most influential contribution is the S2R-Pick framework, a generic deep-learning system that achieves fast, accurate object recognition and localization for industrial robotic bin picking, overcoming challenges posed by textureless and reflective parts common in manufacturing. This work has garnered 52 citations, reflecting its practical significance. Feng also developed SESR, a self-ensembling approach for instance segmentation in auto-store scenarios that eliminates the need for costly manual annotations, demonstrating a commitment to scalable, cost-effective solutions. By pioneering robust sim-to-real pipelines, Feng has advanced the deployment of intelligent robotic systems in logistics and manufacturing, enabling more reliable automation. Their research bridges the gap between computer vision theory and industrial practice, making them a key figure in the field of robotic perception and manipulation.
Research Focus
Key Achievements
Top Papers
- 1
- 2