Xiantong Zhen

University of Amsterdam, Western University, Guangdong University of Petrochemical Technology

Papers

Total Citations

396

H-Index

About

Xiantong Zhen is a leading researcher in computer vision and machine learning, with a primary focus on 3D human pose estimation and scene flow analysis. His most impactful contribution is the comprehensive review "Deep 3D human pose estimation: A review" (2021), which has garnered 383 citations, establishing itself as a foundational reference for researchers working on estimating articulated 3D joint locations from images and video—a critical task for applications in human motion analysis, human-computer interaction, and robotics. Beyond this, Zhen has explored the clinical application of video-based motion analysis, as demonstrated in his work on classifying normal and abnormal status in epileptic patients using human action recognition. More recently, he has advanced the field of 3D scene understanding through innovative deep learning architectures, including the integration of Transformers for deep scene flow learning from point clouds and the development of PVFT-Net, a point-voxel fusion method for self-supervised scene flow estimation. His research consistently bridges fundamental algorithmic development with practical, real-world applications, making significant strides in both human-centric and geometric computer vision.