Song–Hai Zhang

Tsinghua University

Papers

6

Total Citations

67

H-Index

4

About

Song–Hai Zhang is a leading researcher in visual computing and autonomous perception, whose work bridges 3D scene understanding, place recognition, and self-supervised depth estimation. His most impactful contribution, TransLoc3D, introduces adaptive receptive fields for point-cloud-based large-scale place recognition—a critical capability for autonomous driving and robot navigation. This work has accumulated 48 combined citations, reflecting its influence on the field. Zhang has also advanced self-supervised monocular depth estimation with PPEA-Depth, a progressive parameter-efficient adaptation method that tackles the static-scene assumption in dynamic environments, and with MAL, a motion-aware loss framework incorporating temporal and distillation hints for multi-frame depth estimation. His research consistently addresses real-world challenges in mobile robotics, including multi-sensor fusion for 3D environment perception and reconstruction. By developing methods that work robustly in dynamic, unlabeled environments, Zhang is helping to make autonomous systems more reliable and data-efficient. His contributions are essential reading for anyone working in 3D vision, autonomous navigation, or self-supervised learning for robotics.

Research Focus

Key Achievements

4
H-Index
6
Papers
67
Total Citations
11
Avg Citations/Paper
🏆 Most Cited Paper
TransLoc3D: point cloud based large-scale place recognition using adaptive receptive fields
38 citations · 2023
📈 Most Prolific Year: 2024 (2 Papers)
🤝 Key Collaborators: 12
🏛 Institutions: Tsinghua University

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 14 days ago