Haodong Chen
Papers
1
Total Citations
3
H-Index
1
About
Haodong Chen is an emerging researcher specializing in **pedestrian trajectory prediction** and **spatial-temporal deep learning**, with a focus on integrating scene context into graph-based neural network architectures. His most notable work introduces the Scene-Constrained Spatial-Temporal Graph Convolutional Neural Network (Scene-STGCNN), a sophisticated framework that addresses a critical limitation in existing trajectory prediction models: the inability to explicitly model how physical scene environments regulate individual pedestrian motion. Rather than relying on implicit feature merging, Chen's architecture incorporates a dedicated scene-based fine-tuning module alongside spatial-temporal convolution and extrapolation components, enabling more interpretable and precise predictions — particularly in densely crowded scenarios where prior graph attention-based methods struggled. Published in 2023, this work has already accumulated 3 citations, reflecting its early but growing recognition within the computer vision and autonomous systems communities. Chen's contributions are particularly relevant to real-world applications such as autonomous driving, robotics, and smart surveillance, where accurate human motion forecasting in complex environments remains a fundamental challenge. His research represents a promising step toward more explainable and scene-aware trajectory prediction systems.
Research Focus
Key Achievements
Top Papers
- 1