Research map / Detection and segmentation / Open-set and few-shot detection
Reasoning About Human-Object Interactions Through Dual Attention Networks
Tete Xiao, Quanfu Fan, Dan Gutfreund, Mathew Monfort, Aude Oliva, Bolei Zhou
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 |
|---|---|---|---|
| add dual attention (ablation) | TRN-4 · Something-Something V2 val (TRN-4) | action top-1: 44.6 -> 46.6 | Improved |
| add dual attention (ablation) | TRN-4 · Something-Something V2 val (TRN-4) | object top-1: 58.2 -> 60.1 | Improved |
| use dual attention instead of multi-task learning (ablation) | TRN-4 · Something-Something V2 val (TRN-4) | action top-1: 45.7 -> 46.6 | Improved |
| use action/object priors instead of self attention (ablation) | Dual Attention · Something-Something V2 val (TRN-4) | action top-1: 45.3 -> 46.6 | Improved |
| use ground-truth object labels as the attention prior (ablation) | TRN-4 · Something-Something V2 val (TRN-4) | action top-1: 44.6 -> 50.2 | Improved |
| use ground-truth action labels as the attention prior (ablation) | TRN-4 · Something-Something V2 val (TRN-4) | object top-1: 58.2 -> 67.0 | Improved |
| add dual attention | TSN · Something-Something V2 val | top-1: 41.1 -> 42.1 | Improved |
| add dual attention | TSM · Something-Something V2 val | top-1: 54.0 -> 55.0 | Improved |
| add dual attention | TRN RGB · Something-Something V2 val | top-1: 48.8 -> 51.6 | Improved |
| add dual attention | TRN RGB+Flow · Something-Something V2 val | top-1: 55.5 -> 58.4 | Improved |
| add dual attention | TRN RGB+Flow · Something-Something V2 test | top-1: 56.2 -> 60.1 | Improved |
| use dual attention instead of self attention for localization (ablation) | attention model · Something-Something V2 val subset | AP@IoU0.5: 26.4 -> 33.7 | Improved |
Closest papers by meaning
- Attentional Pooling for Action Recognition (NeurIPS 2017 · similarity 0.88)
- Object Level Visual Reasoning in Videos (ECCV 2018 · similarity 0.87)
- Motion Guided Attention Fusion To Recognize Interactions From Videos (ICCV 2021 · similarity 0.87)
- Pose-Aware Multi-Level Feature Network for Human Object Interaction Detection (ICCV 2019 · similarity 0.87)
- Where and Why Are They Looking? Jointly Inferring Human Attention and Intentions in Complex Tasks (CVPR 2018 · similarity 0.87)
- Learning from Observer Gaze: Zero-Shot Attention Prediction Oriented by Human-Object Interaction Recognition (CVPR 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.
