Research map / Detection and segmentation / Open-set and few-shot detection
Streamlined Open-Vocabulary Human-Object Interaction Detection
Chang Sun, Dongliang Liao, Changxing Ding
Read the paperCheck 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 |
|---|---|---|---|
| feed interaction queries with image tokens through frozen vision head (Semantic Bootstrapping) (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 21.82 -> 23.28 | Improved |
| feed interaction queries with image tokens through frozen vision head (Semantic Bootstrapping) (ablation) | SL-HOI · SWiG-HOI (Unseen) | mAP: 16.55 -> 18.09 | Improved |
| add Hierarchical Refinement cross-attention block (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 23.28 -> 24.67 | Improved |
| add Hierarchical Refinement cross-attention block (ablation) | SL-HOI · SWiG-HOI (Unseen) | mAP: 18.09 -> 19.04 | Improved |
| let interaction queries influence image tokens (no attention mask) (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 24.01 -> 24.67 | Improved |
| process queries in frozen head instead of multi-scale late fusion decoder (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 21.77 -> 23.28 | Improved |
| add backbone tokens to late fusion decoder (multi-scale) (ablation) | SL-HOI · SWiG-HOI (Unseen) | mAP: 16.55 -> 15.73 | Hurt |
| add backbone tokens to late fusion decoder (multi-scale) (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 21.82 -> 21.77 | No clear effect |
| use 2 detection adapter encoder layers rather than 0 (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 23.95 -> 24.67 | Improved |
| use 2 rather than 5 detection adapter encoder layers (ablation) | SL-HOI · SWiG-HOI (Full) | mAP: 23.29 -> 24.67 | Improved |
| SL-HOI vs BC-HOI on unseen HICO-DET open-vocab | SL-HOI · HICO-DET open-vocabulary (Unseen) | mAP: 42.31 -> 40.53 | Hurt |
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.
- SGC-Net — SGC-Net: Stratified Granular Comparison Network for Open-Vocabulary HOI Detection (CVPR 2025) · better on 2 of 2 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
- Exploring the Potential of Large Foundation Models for Open-Vocabulary HOI Detection (CVPR 2024 · similarity 0.95)
- Open-World Human-Object Interaction Detection via Multi-modal Prompts (CVPR 2024 · similarity 0.92)
- SGC-Net: Stratified Granular Comparison Network for Open-Vocabulary HOI Detection (CVPR 2025 · similarity 0.91)
- Discovering Syntactic Interaction Clues for Human-Object Interaction Detection (CVPR 2024 · similarity 0.91)
- Toward Open-Set Human Object Interaction Detection (AAAI 2024 · similarity 0.91)
- Exploring Pose-Aware Human-Object Interaction via Hybrid Learning (CVPR 2024 · similarity 0.90)
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.
