SGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction

@article{Shi2021SGCNSparseGC,
  title={SGCN:Sparse Graph Convolution Network for Pedestrian Trajectory Prediction},
  author={Liushuai Shi and Le Wang and Chengjiang Long and Sanping Zhou and Mo Zhou and Zhenxing Niu and Gang Hua},
  journal={2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2021},
  pages={8990-8999}
}
  • Liushuai ShiLe Wang G. Hua
  • Published 4 April 2021
  • Computer Science
  • 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Pedestrian trajectory prediction is a key technology in autopilot, which remains to be very challenging due to complex interactions between pedestrians. However, previous works based on dense undirected interaction suffer from modeling superfluous interactions and neglect of trajectory motion tendency, and thus inevitably result in a considerable deviance from the reality. To cope with these issues, we present a Sparse Graph Convolution Network (SGCN) for pedestrian trajectory prediction… 

Figures and Tables from this paper

Complementary Attention Gated Network for Pedestrian Trajectory Prediction

A complementary attention gated network (CAGN) is proposed for pedestrian trajectory prediction, in which a dual-path architecture including normal and inverse attention is proposed to capture both frequent and peculiar modals in spatial and temporal patterns, respectively.

OA-STGCN: An Output Anchoring-based Graph Convolutional Network for Human Trajectory Prediction

An Output Anchoring based Spatio-Temporal Graph Convolution Network (OA-STGCN) for human trajectory prediction, which considers the subjective faces of pedestrian generating target trajectory, includes the Time-Extrapolator Convolutional Neural Network (TXPCNN) and is followed by a simple and effective Attention Anchoring Block (AAB).

Pose Forecasting in Industrial Human-Robot Collaboration

A new Separable-Sparse Graph Convolutional Network (SeS-GCN) for pose forecasting, which bottlenecks the interaction of the spatial, temporal and channel-wise dimensions in GCNs, and it learns sparse adjacency matrices by a teacher-student framework.

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

Multiclass-SGCN is proposed, 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.

ATPFL: Automatic Trajectory Prediction Model Design under Federated Learning Framework

In ATPFL, an effective TP search space is built by analyzing and summarizing the existing works, and a relation-sequence-aware search strategy is designed, realizing the automatic design of the TP model.

SSAGCN: Social Soft Attention Graph Convolution Network for Pedestrian Trajectory Prediction

A new prediction model named Social Soft Attention Graph Convolution Network (SSAGCN) which aims to simultaneously handle social interactions among pedestrians and scene interactions between pedestrians and environments and successfully obtains socially and physically acceptable predicted trajectories.

STGlow: A Flow-based Generative Framework with Dual Graphormer for Pedestrian Trajectory Prediction

This paper proposes a novel generative flow based framework with dual graphormer combining with the graph structure to more adequately model the temporal dependencies and the mutual spatial interactions for pedestrian trajectory prediction (STGlow).

EPC: a video analytics system with efficient edge-side privacy control

Evaluation results show that EPC can prevent privacy leakage under visual data attack with 95% video analytics accuracy and a 4x speedup compared to existing privacy control mechanisms.

Expanding the Deployment Envelope of Behavior Prediction via Adaptive Meta-Learning

The approach leverages recent advances in meta-learning, Bayesian regres- sion, to augment existing behavior prediction models with an adaptive layer that enables efficient domain transfer via offline tuning, online adaptation, or both and can adapt to a variety of unseen environments.

D2-TPred: Discontinuous Dependency for Trajectory Prediction under Traffic Lights

A trajectory prediction approach with respect to traffic lights, D2-TPred, which uses a spatial dynamic interaction graph (SDG) and a behavior dependency graph (BDG) to handle the problem of discontinuous dependency in the spatial-temporal space.

References

SHOWING 1-10 OF 48 REFERENCES

Attention is All you Need

A new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely is proposed, which generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.

Semi-Supervised Classification with Graph Convolutional Networks

A scalable approach for semi-supervised learning on graph-structured data that is based on an efficient variant of convolutional neural networks which operate directly on graphs which outperforms related methods by a significant margin.

Spatio-Temporal Graph Transformer Networks for Pedestrian Trajectory Prediction

STAR is presented, a Spatio-Temporal grAph tRansformer framework, which tackles trajectory prediction by only attention mechanisms, and achieves state-of-the-art performance on 5 commonly used real-world pedestrian prediction datasets.

Recursive Social Behavior Graph for Trajectory Prediction

A novel insight of group-based social interaction model to explore relationships among pedestrians is presented and state-of-the-art methods on ETH and UCY dataset are surpassed for 11.1% in ADE and 10.8% in FDE in average.

Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction

The Social Spatio-Temporal Graph Convolutional Neural Network (Social-STGCNN), which substitutes the need of aggregation methods by modeling the interactions as a graph, and is data efficient, and exceeds previous state of the art on the ADE metric with only 20% of the training data.

Social-BiGAT: Multimodal Trajectory Forecasting using Bicycle-GAN and Graph Attention Networks

A graph-based generative adversarial network that generates realistic, multimodal trajectory predictions by better modelling the social interactions of pedestrians in a scene and achieves state-of-the-art performance comparing it to several baselines on existing trajectory forecasting benchmarks.

Motion Planning and Control of Autonomous Driving Intelligence System Based on Risk Potential Optimization Framework

This study proposes a motion planning and control system based on collision risk potential prediction characteristics of experienced drivers that optimizing the potential field function in the framework of optimal control theory, the desired yaw rate and the desired longitudinal deceleration are theoretically calculated.

Large Scale GAN Training for High Fidelity Natural Image Synthesis

It is found that applying orthogonal regularization to the generator renders it amenable to a simple "truncation trick," allowing fine control over the trade-off between sample fidelity and variety by reducing the variance of the Generator's input.

SoPhie: An Attentive GAN for Predicting Paths Compliant to Social and Physical Constraints

SoPhie is presented; an interpretable framework based on Generative Adversarial Network (GAN), which leverages two sources of information, the path history of all the agents in a scene, and the scene context information, using images of the scene.

Social GAN: Socially Acceptable Trajectories with Generative Adversarial Networks

A recurrent sequence-to-sequence model observes motion histories and predicts future behavior, using a novel pooling mechanism to aggregate information across people, and outperforms prior work in terms of accuracy, variety, collision avoidance, and computational complexity.