Research map / Detection and segmentation / Open-set and few-shot detection
No-Frills Human-Object Interaction Detection: Factorization, Layout Encodings, and Training Techniques
Tanmay Gupta, Alexander Schwing, Derek Hoiem
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 |
|---|---|---|---|
| raise negative to positive ratio in mini-batches from 10 to 1000 (ablation) | Det+App+Box · HICO-Det (Full, Det+App+Box) | mAP: 13.40 -> 16.96 | Improved |
| raise negative to positive ratio from 1000 to 1500 (ablation) | Det+App+Box · HICO-Det (Full, Det+App+Box) | mAP: 16.96 -> 16.62 | No clear effect |
| add indicator terms during training (ablation) | Det+App+Box · HICO-Det (Full, Det+App+Box) | mAP: 15.93 -> 16.96 | Improved |
| train with HOI classification loss instead of interaction loss (ablation) | Det+App+Box · HICO-Det (Full, Det+App+Box) | mAP: 15.89 -> 16.96 | Improved |
| add box configuration factor to detection terms (ablation) | Det · HICO-Det (Full) | mAP: 8.32 -> 12.54 | Improved |
| add appearance factor to detection terms (ablation) | Det · HICO-Det (Full) | mAP: 8.32 -> 15.74 | Improved |
| add pose factor to detection terms (ablation) | Det · HICO-Det (Full) | mAP: 8.32 -> 11.09 | Improved |
| combine human and object appearance instead of either alone (ablation) | Det + App · HICO-Det (Full) | mAP: 11.12 -> 15.74 | Improved |
| add pose factor (ablation) | Det + App + Box · HICO-Det (Full) | mAP: 16.96 -> 17.18 | Improved |
| add pose factor (ablation) | Det + App · HICO-Det (Full) | mAP: 15.74 -> 15.50 | No clear effect |
| add pose factor (ablation) | Det + Box · HICO-Det (Full) | mAP: 12.54 -> 14.49 | Improved |
| add appearance factor (ablation) | Det + Box + Pose · HICO-Det (Full) | mAP: 14.49 -> 17.18 | Improved |
Closest papers by meaning
- Learning Human-Object Interaction Detection Using Interaction Points (CVPR 2020 · similarity 0.91)
- Exploring Pose-Aware Human-Object Interaction via Hybrid Learning (CVPR 2024 · similarity 0.89)
- Learning Human-Object Interaction as Groups (NeurIPS 2025 · similarity 0.88)
- Disentangled Pre-training for Human-Object Interaction Detection (CVPR 2024 · similarity 0.88)
- Orchestrating the Symphony of Prompt Distribution Learning for Human-Object Interaction Detection (AAAI 2025 · similarity 0.87)
- Toward Open-Set Human Object Interaction Detection (AAAI 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.
