Deming Zhai

Papers

Total Citations

H-Index

About

Deming Zhai is a rising researcher whose work is shaping the future of 3D vision and spatial intelligence. His primary research areas span point cloud processing, 6D object tracking, and self-supervised learning for 3D data. Zhai’s most significant contribution is his pioneering work on **self-supervised arbitrary-scale implicit point clouds upsampling**, published in 2023. This method addresses a critical bottleneck in 3D sensing: generating dense, uniform point clouds from sparse LiDAR inputs without requiring costly labeled data. By enabling high-fidelity upsampling, his technique has direct implications for autonomous driving, robotics, and AR/VR, where sensor sparsity often limits performance. The paper has already garnered **17 citations**, signaling strong interest from the community. More recently, Zhai has pushed boundaries with **Zero6DOT** (2025), a zero-shot approach for 6D object pose tracking using only monocular RGB video. This work eliminates the need for CAD models or multi-modal sensors, making robust tracking accessible for real-world robotic manipulation and virtual reality. His ability to solve practical, data-hungry problems with self-supervised and zero-shot paradigms marks him as a forward-thinking innovator in 3D computer vision.