Yongzhi Su

About

Yongzhi Su is a leading researcher in computer vision and augmented reality, with a focus on 6-degree-of-freedom (6DoF) object pose estimation and its applications in interactive systems. His work bridges the gap between deep learning and practical AR/VR and robotic manipulation, addressing the critical challenge of accurately determining an object’s position and orientation from visual data. Su’s most cited paper, “Deep Multi-state Object Pose Estimation for Augmented Reality Assembly” (2019, 100 citations), pioneered neural network approaches for estimating the states of objects with multiple movable parts, enabling more realistic and functional AR assembly guidance. He further advanced the field with “SynPo-Net” (2021, 19 citations), which introduced a fast and accurate CNN-based method for 6DoF pose estimation using purely synthetic training data—a significant step toward reducing the need for labor-intensive real-world annotation. Su’s contributions are vital for industries ranging from manufacturing to entertainment, where precise object interaction is essential. His work on multi-state estimation and synthetic data training has been widely recognized for its impact on making AR and robotics more robust and scalable.

Research Focus

Key Achievements

Top Papers

1. 1
   Deep Multi-state Object Pose Estimation for Augmented Reality Assembly

   100 citations · 2019
2. 2
   SynPo-Net—Accurate and Fast CNN-Based 6DoF Object Pose Estimation Using Synthetic Training

   19 citations · 2021

Key Collaborators

Contact & Links