Yuzhuo Han
Papers
1
Total Citations
54
H-Index
1
About
Yuzhuo Han is a leading researcher in computer vision and autonomous driving, with a primary focus on semantic segmentation—a critical perception task that classifies every pixel in an image for self-driving cars and robotics. His most cited work, "Importance-Aware Semantic Segmentation in Self-Driving with Discrete Wasserstein Training" (2020, 54 citations), introduces a novel discrete Wasserstein training framework that moves beyond conventional cross-entropy loss. By incorporating importance-aware learning, Han’s method significantly improves mean Intersection-over-Union (mIoU) performance, addressing the challenge of class imbalance and edge-case accuracy in real-world driving scenes. This contribution has been widely recognized for enhancing the reliability of perception systems in safety-critical environments. Han’s research bridges theoretical advances in optimal transport with practical deployment needs, earning him citations from top venues in autonomous systems and robotics. His work continues to influence the development of more robust and efficient deep learning models for scene understanding, making him a notable figure in the intersection of machine learning and autonomous navigation.
Research Focus
Key Achievements
Top Papers
- 1