Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embedding

Ruochen Li, Stamos Katsigiannis and Hubert P. H. Shum
Proceedings of the 2022 IEEE International Conference on Image Processing (ICIP), 2022

Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embedding

Abstract

Trajectory prediction of road users in real-world scenarios is challenging because their movement patterns are stochastic and complex. Previous pedestrian-oriented works have been successful in modelling the complex interactions among pedestrians, but fail in predicting trajectories when other types of road users are involved (e.g., cars, cyclists, etc.), because they ignore user types. Although a few recent works construct densely connected graphs with user label information, they suffer from superfluous spatial interactions and temporal dependencies. To address these issues, we propose Multiclass-SGCN, a sparse graph convolution network based approach for multi-class trajectory prediction that takes into consideration velocity and agent label information and uses a novel interaction mask to adaptively decide the spatial and temporal connections of agents based on their interaction scores. The proposed approach significantly outperformed state-of-the-art approaches on the Stanford Drone Dataset, providing more realistic and plausible trajectory predictions.

Downloads

YouTube

Citations

BibTeX

@inproceedings{li22multiclasssgcn,
 author={Li, Ruochen and Katsigiannis, Stamos and Shum, Hubert P. H.},
 booktitle={Proceedings of the 2022 IEEE International Conference on Image Processing},
 series={ICIP '22},
 title={Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embedding},
 year={2022},
 publisher={IEEE},
 location={Bordeaux, France},
}

RIS

TY  - CONF
AU  - Li, Ruochen
AU  - Katsigiannis, Stamos
AU  - Shum, Hubert P. H.
T2  - Proceedings of the 2022 IEEE International Conference on Image Processing
TI  - Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embedding
PY  - 2022
PB  - IEEE
ER  - 

Plain Text

Ruochen Li, Stamos Katsigiannis and Hubert P. H. Shum, "Multiclass-SGCN: Sparse Graph-based Trajectory Prediction with Agent Class Embedding," in ICIP '22: Proceedings of the 2022 IEEE International Conference on Image Processing, Bordeaux, France, IEEE, 2022.

Similar Research

Qianhui Men and Hubert P. H. Shum, "PyTorch-based Implementation of Label-aware Graph Representation for Multi-class Trajectory Prediction", Software Impacts (SIMPAC), 2021
Ben Rainbow, Qianhui Men and Hubert P. H. Shum, "Semantics-STGCNN: A Semantics-guided Spatial-Temporal Graph Convolutional Network for Multi-class Trajectory Prediction", Proceedings of the 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2021

 

 
 

Last updated on 22 September 2022, RSS Feeds