Hongsong Wang

National University of Singapore, Southeast University

Papers

Total Citations

H-Index

About

Hongsong Wang is a researcher specializing in human motion analysis, motion forecasting, and skeleton-based action understanding — areas that sit at the intersection of computer vision, deep learning, and intelligent systems. His work addresses fundamental challenges in modeling and predicting human behavior from structural representations of the body. Among his notable contributions is his 2021 paper on velocity-to-velocity human motion forecasting, which garnered 10 citations and introduced innovative approaches to predicting fluid, realistic human movement sequences — a critical problem for robotics and animation applications. More recently, his 2025 work on a Foundation Model for Skeleton-Based Human Action Understanding reflects his forward-looking research agenda, exploring scalable, generalizable frameworks for action recognition that leverage skeleton data as a modality- and device-agnostic representation. This work holds particular promise for humanoid robot control and human-robot interaction. Wang's research contributes to the growing field of intelligent motion perception, where robust understanding of human actions is essential for building responsive, context-aware AI systems. His trajectory from motion forecasting to foundation models signals a researcher steadily expanding the scope and ambition of his contributions to human-centric AI.