Yonghao Dang

Beijing University of Posts and Telecommunications

Papers

Total Citations

H-Index

About

Yonghao Dang is an emerging researcher whose work sits at the intersection of computer vision, human action recognition, and embodied robotics. His research primarily focuses on 3D skeleton-based action recognition and the integration of multimodal large language models into robotic systems. Dang's most recognized contribution, "DWnet: Deep-Wide Network for 3D Action Recognition" (2020), has accumulated 20 citations and demonstrates his early commitment to designing architectures that effectively capture spatiotemporal features from skeletal data. Building on this foundation, his 2025 work on generically contrastive spatiotemporal representation enhancement further advances the field by strengthening how models learn discriminative features for skeleton-based action understanding. More recently, Dang has pushed into cutting-edge robotics research with Quart-Online, a latency-free multimodal large language model framework designed specifically for quadruped robots — tackling the practical challenge of real-time vision-language-action deployment without sacrificing model performance. Though early in his career, Dang's trajectory reflects a researcher steadily bridging the gap between human motion understanding and intelligent embodied systems, making his work increasingly relevant to both the computer vision and autonomous robotics communities.