Research map / Detection and segmentation / Open-set and few-shot detection
VSGNet: Spatial Attention Network for Detecting Human Object Interactions Using Graph Convolutions
Oytun Ulutan, A S M Iftekhar, B. S. Manjunath
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 the Spatial attention branch to the Visual branch (ablation) | VSGNet · V-COCO test, Scenario 1 | mAP: 47.3 -> 50.33 | Improved |
| add the Spatial attention branch to the Visual branch (ablation) | VSGNet · V-COCO test, Scenario 2 | mAP: 52.15 -> 55.32 | Improved |
| add the Graph branch to the Visual branch without Spatial attention (ablation) | VSGNet · V-COCO test, Scenario 1 | mAP: 47.3 -> 48.19 | Improved |
| add the Graph branch to the Visual branch without Spatial attention (ablation) | VSGNet · V-COCO test, Scenario 2 | mAP: 52.15 -> 53.12 | Improved |
| add the Graph branch to the Visual+Spatial model (ablation) | VSGNet · V-COCO test, Scenario 1 | mAP: 50.33 -> 51.76 | Improved |
| add the Graph branch to the Visual+Spatial model (ablation) | VSGNet · V-COCO test, Scenario 2 | mAP: 55.32 -> 57.03 | Improved |
| replace Resnet-152 with VGG-19 as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 48.37 | Hurt |
| replace Resnet-152 with InceptionV3 as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 49.39 | Hurt |
| replace Resnet-152 with SqueezeNet as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 43.4 | Hurt |
| replace Resnet-152 with Resnet34 as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 50.88 | Hurt |
| replace Resnet-152 with Resnet50 as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 51.01 | Hurt |
| replace Resnet-152 with Resnet101 as the backbone (ablation) | VSGNet · V-COCO, Scenario 1 | mAP: 51.76 -> 50.01 | Hurt |
Later papers that compared with this method
6 later accepted papers report VSGNet as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Exploiting Scene Graphs for Human-Object Interaction Detection (ICCV 2021 · better on 2 of 2 shared results)
- Polysemy Deciphering Network for Robust Human–Object Interaction Detection (IJCV 2021 · better on 2 of 2 shared results)
- DIRV: Dense Interaction Region Voting for End-to-End Human-Object Interaction Detection (AAAI 2021 · better on 1 of 1 shared result)
- Spatially Conditioned Graphs for Detecting Human-Object Interactions (ICCV 2021 · better on 1 of 1 shared result)
- HOI Analysis: Integrating and Decomposing Human-Object Interaction (NeurIPS 2020 · better on 2 of 2 shared results)
- Polysemy Deciphering Network for Human-Object Interaction Detection (ECCV 2020 · better on 1 of 1 shared result)
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
- Glance and Gaze: Inferring Action-Aware Points for One-Stage Human-Object Interaction Detection (CVPR 2021 · similarity 0.87)
- Spatially Conditioned Graphs for Detecting Human-Object Interactions (ICCV 2021 · similarity 0.87)
- Learning Human-Object Interaction Detection Using Interaction Points (CVPR 2020 · similarity 0.87)
- Contextual Heterogeneous Graph Network for Human-Object Interaction Detection (ECCV 2020 · similarity 0.87)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.85)
- Exploring Self- and Cross-Triplet Correlations for Human-Object Interaction Detection (AAAI 2024 · similarity 0.85)
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.
