Chengzhong Wu

Papers

Total Citations

H-Index

About

Chengzhong Wu is a rising researcher in computer vision and robotics, whose work focuses on enabling machines to perceive and interact with their environment with unprecedented precision. His primary research areas include 6-DoF object pose estimation and robotic grasping, where he tackles the fundamental challenge of bridging perception and action for unseen objects. Wu’s most notable contribution is the **PoseDiffusion** framework, a coarse-to-fine diffusion model that achieves robust 6-DoF pose estimation for objects never seen during training—a critical capability for industrial automation. This work has already garnered 6 citations since its 2024 publication, signaling strong early impact. Complementing this, his **Vim-Grasp** system introduces a novel Mamba-based architecture for generating multi-scale grasping gestures, directly addressing the robustness and accuracy gaps in complex environments. By leveraging state-space models for gesture generation, Wu pushes beyond traditional transformer-based approaches, offering a more efficient pathway for real-time robotic manipulation. His research is particularly notable for its practical orientation: both PoseDiffusion and Vim-Grasp are designed to operate without object-specific priors, making them directly applicable to dynamic, unstructured settings. As his citation counts grow, Wu is establishing himself as a key voice in the next generation of vision-for-robotics research.