Liusheng Huang
Papers
3
Total Citations
102
H-Index
3
About
Liusheng Huang is a leading researcher in 3D computer vision, with a primary focus on advancing object detection and adversarial robustness for autonomous driving and robotics. His most impactful contribution is the development of ZoomNet, a pioneering framework for stereo imagery-based 3D object detection that addresses the critical challenge of accurately estimating the pose of distant and occluded objects. This work, which has garnered 81 citations, introduces a part-aware adaptive zooming mechanism that significantly enhances detection precision in complex, real-world environments. Beyond detection, Huang has made notable strides in the security of 3D deep learning models. His research on shape-prior-guided attacks on 3D point clouds (17 citations) demonstrates how to craft sparser yet more effective adversarial perturbations, exposing vulnerabilities in classification models used in robots and drones. By tackling both the accuracy and robustness of 3D perception systems, Huang’s work is instrumental in building safer, more reliable autonomous technologies.
Research Focus
Key Achievements
Top Papers
- 1ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection81 citations · 2020
- 2Shape Prior Guided Attack: Sparser Perturbations on 3D Point Clouds17 citations · 2022
- 3