Xijun Wang

University of Maryland, College Park, China Southern Power Grid (China)

Papers

Total Citations

H-Index

About

Xijun Wang is a rising researcher at the intersection of computer vision, robotics, and human-machine interaction, with a primary focus on aerial video analysis and teleoperated systems. His most impactful contribution is in aerial video action recognition, where he pioneered the AZTR framework (2023, 19 citations)—a novel approach that integrates auto zoom and temporal reasoning to enable robust action recognition from UAV-captured footage on edge devices. This work addresses the critical challenge of identifying and scaling human targets in dynamic, low-resolution aerial views. Building on this, Wang introduced Soft Conditional Prompt Learning (SCP, 2024, 3 citations), a prompt-based method that enhances model focus on agent-specific instructions, pushing the boundaries of efficient aerial action prediction. In robotics, he has advanced teleoperation by designing a mixed reality-based electric power live line working robot (2024, 4 citations), which combines graphic and haptic feedback for enhanced telepresence and operational safety. Wang’s contributions are particularly notable for their practical deployment focus—bridging real-time edge computing with complex visual reasoning—and his work is gaining traction in both autonomous systems and human-robot collaboration communities.