Research map / Video / Action recognition
Graph-Based High-Order Relation Modeling for Long-Term Action Recognition
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace the I3D baseline with GHRM | Breakfast long-term action recognition with a Kinetics-pretrained backbone · Breakfast | mAP: 63.14 -> 65.86 | Helped |
| replace the I3D baseline with GHRM | Breakfast long-term action recognition with a Kinetics-pretrained backbone · Breakfast | Acc: 69.45 -> 75.49 | Helped |
| add GHRM to the fine-tuned I3D backbone | Breakfast long-term action recognition · Breakfast | mAP: 61.19 -> 73.89 | Helped |
| add GHRM to I3D | Charades sub-action recognition · Charades | mAP: 37.8 -> 38.3 | Helped |
| add GHRM to the Kinetics-pretrained I3D backbone | MultiThumos sub-action recognition · MultiThumos | mAP: 74.79 -> 79.89 | Helped |
| use 32 rather than 8 graphs (ablation) | GHRM · Breakfast | mAP: 61.2 -> 65.9 | Helped |
| increase the graph count from 32 to 64 (ablation) | GHRM · Breakfast | mAP: 65.9 -> 65.7 | Hurt |
| increase the graph count from 32 to 128 (ablation) | GHRM · Breakfast | mAP: 65.9 -> 63.8 | Hurt |
| use 64 rather than 8 graphs (ablation) | GHRM · Charades | mAP: 36.5 -> 38.3 | Helped |
| increase the graph count from 64 to 128 (ablation) | GHRM · Charades | mAP: 38.3 -> 38.1 | Hurt |
| add the Temporal-GHRM branch (ablation) | Semantic-GHRM on Breakfast · Breakfast | mAP: 56.4 -> 65.9 | Helped |
| add the Semantic-GHRM branch (ablation) | Temporal-GHRM on Breakfast · Breakfast | mAP: 62.9 -> 65.9 | Helped |
Closest papers by meaning
- Object-Relation Reasoning Graph for Action Recognition (CVPR 2022 · similarity 0.89)
- Temporal Relational Modeling with Self-Supervision for Action Segmentation (AAAI 2021 · similarity 0.88)
- I Know the Relationships: Zero-Shot Action Recognition via Two-Stream Graph Convolutional Networks and Knowledge Graphs (AAAI 2019 · similarity 0.87)
- Open-Vocabulary Video Relation Extraction (AAAI 2024 · similarity 0.86)
- Spatio-Temporal Fusion for Human Action Recognition via Joint Trajectory Graph (AAAI 2024 · similarity 0.86)
- Dynamic Semantic-Based Spatial Graph Convolution Network for Skeleton-Based Human Action Recognition (AAAI 2024 · similarity 0.86)
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.