Zeyun Zhong

Papers

Total Citations

H-Index

About

Zeyun Zhong is a researcher advancing the frontiers of human action recognition and anticipation, with a focus on generalization and real-world applicability. Their work addresses critical challenges in surveillance, robotics, and automotive safety, particularly occupant monitoring. Zhong’s major contributions include pioneering unsupervised 3D skeleton-based action recognition using cross-attention mechanisms with conditioned generation capabilities, a method that enhances model robustness without labeled data. They also introduced SynthAct, a framework leveraging synthetic data to augment training sets for human-robot collaboration in privacy-sensitive household environments. Additionally, Zhong authored a comprehensive survey on deep learning techniques for action anticipation, covering applications from autonomous driving to human-robot interaction. With their most cited paper garnering 7 citations, Zhong’s research is gaining traction for its innovative approach to data scarcity and domain generalization. Their work on synthetic data generation and cross-attention models represents a significant step toward more adaptable and privacy-preserving AI systems, making them a rising voice in the field of computer vision and human activity analysis.