Shi‐Min Hu
Papers
6
Total Citations
72
H-Index
5
About
Shi-Min Hu is a computer scientist whose research spans 3D scene reconstruction, multi-sensor fusion, semantic segmentation, and computational design, with growing contributions to robotics and autonomous systems. Among his most recognized contributions is a noise-resilient framework for reconstructing large-scale indoor panoramas and 3D scenes using unsynchronized RGB-D cameras mounted on robotic platforms, which has accumulated 25 citations and addresses a longstanding challenge of sensor synchronization in real-world environments. His work on HeteroFusion (20 citations) advanced dense scene reconstruction by integrating heterogeneous sensor data in real time, overcoming the limitations of single-camera approaches that struggle with sparse geometric features. Hu has also pushed the boundaries of video understanding through LinkNet, a 2D-3D linked multi-modal network for real-time semantic segmentation of RGB-D video streams. Beyond perception, his interdisciplinary reach extends to computational design—notably a pioneering tool for creating transforming pop-up books—and to robotics software security, including fuzzing techniques for Robot Operating System programs. His most recent work on robotic strawberry harvesting reflects a compelling application of path-planning methods to agricultural automation. Across these diverse domains, Hu consistently bridges theoretical innovation with practical engineering impact.
Research Focus
Key Achievements
Top Papers
- 1
- 2HeteroFusion: Dense Scene Reconstruction Integrating Multi-Sensors20 citations · 2019
- 3
- 4Computational Design of Transforming Pop-up Books6 citations · 2018
- 5
- 6