Research map / Graph learning / Knowledge and temporal graphs
SGCN: Sparse Graph Convolution Network for Pedestrian Trajectory Prediction
Liushuai Shi, Le Wang, Chengjiang Long, Sanping Zhou, Mo Zhou, Zhenxing Niu, Gang Hua
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add Motion Tendency to Sparse Directed Interaction (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.62 -> 0.37 | Improved |
| Add Motion Tendency to Sparse Directed Interaction (ablation) | SGCN · ETH/UCY (AVG) | FDE: 1.19 -> 0.65 | Improved |
| Use Zero-Softmax rather than Softmax for sparse adjacency normalization (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.41 -> 0.37 | Improved |
| Use Zero-Softmax rather than Softmax for sparse adjacency normalization (ablation) | SGCN · ETH/UCY (AVG) | FDE: 0.76 -> 0.65 | Improved |
| Add Sparse Directed Interaction to Motion Tendency (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.66 -> 0.37 | Improved |
| Add Sparse Directed Interaction to Motion Tendency (ablation) | SGCN · ETH/UCY (AVG) | FDE: 1.28 -> 0.65 | Improved |
| Set interaction threshold ξ=1, eliminating interactions (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.37 -> 0.49 | Hurt |
| Set interaction threshold ξ=0.75 for very sparse interactions (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.37 -> 0.40 | Hurt |
| Set interaction threshold ξ=0.25 for relatively dense interactions (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.37 -> 0.42 | Hurt |
| Set interaction threshold ξ=0 for dense interactions (ablation) | SGCN · ETH/UCY (AVG) | ADE: 0.37 -> 0.45 | Hurt |
| Use sparse directed interaction and motion tendency instead of the strongest reported baselines | SGCN · ETH/UCY (AVG) | ADE: 0.41 -> 0.37 | Improved |
| Use sparse directed interaction and motion tendency instead of the strongest reported baselines | SGCN · ETH/UCY (AVG) | FDE: 0.75 -> 0.65 | Improved |
Later papers that compared with this method
4 later accepted papers report SGCN as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- Self-Explainable Graph Transformer for Link Sign Prediction (AAAI 2025 · better on 5 of 7 shared results)
- DropEdge not Foolproof: Effective Augmentation Method for Signed Graph Neural Networks (NeurIPS 2024 · better on 6 of 6 shared results)
- A Set of Control Points Conditioned Pedestrian Trajectory Prediction (AAAI 2023 · better on 3 of 3 shared results)
- Learning Pedestrian Group Representations for Multi-modal Trajectory Prediction (ECCV 2022 · better on 4 of 4 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Social-STGCNN — Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction (CVPR 2020) · better on 4 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction (CVPR 2020 · similarity 0.92)
- A Set of Control Points Conditioned Pedestrian Trajectory Prediction (AAAI 2023 · similarity 0.91)
- A Unified Environmental Network for Pedestrian Trajectory Prediction (AAAI 2024 · similarity 0.88)
- GCNext: Towards the Unity of Graph Convolutions for Human Motion Prediction (AAAI 2024 · similarity 0.87)
- SocialCVAE: Predicting Pedestrian Trajectory via Interaction Conditioned Latents (AAAI 2024 · similarity 0.87)
- MS-TIP: Imputation Aware Pedestrian Trajectory Prediction (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
