Research map / Detection and segmentation / CNN features
End-to-End Human Object Interaction Detection With HOI Transformer
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 |
|---|---|---|---|
| Use HOI Transformer with ResNet-101 instead of the strongest one-stage baseline | HICO-DET · HICO-DET test, Default Full | AP: 21.73 -> 26.61 | Helped |
| Use HOI Transformer with ResNet-101 instead of the strongest prior method on Default Full | HICO-DET · HICO-DET test, Default Full | AP: 24.53 -> 26.61 | Helped |
| Use HOI Transformer with ResNet-101 instead of the strongest one-stage baseline | HICO-DET · HICO-DET test, Default Rare | AP: 13.78 -> 19.15 | Helped |
| Use HOI Transformer with ResNet-101 instead of the strongest prior method on Default Rare | HICO-DET · HICO-DET test, Default Rare | AP: 21.17 -> 19.15 | Hurt |
| Use HOI Transformer with ResNet-101 instead of the strongest baseline | HICO-DET · HICO-DET test, Known Object Full | AP: 27.98 -> 29.13 | Helped |
| Use HOI Transformer instead of the top-scoring prior baseline | V-COCO · V-COCO test | AP_role: 53.2 -> 52.9 | Hurt |
| Increase object classification loss weight to 2.0 (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 17.49 | Helped |
| Increase interaction classification loss weight to 2.0 (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 17.85 | Helped |
| Increase localization weight in matching cost to 2.0 (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 16.20 | Hurt |
| Increase classification weight in matching cost to 2.0 (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 18.46 | Helped |
| Add random crop augmentation (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 22.05 | Helped |
| Add multi-scale training (ablation) | HOI Transformer · HICO-DET ablation, Default Full | AP: 16.97 -> 21.26 | Helped |
Closest papers by meaning
- Exploring Structure-Aware Transformer Over Interaction Proposals for Human-Object Interaction Detection (CVPR 2022 · similarity 0.94)
- HOTR: End-to-End Human-Object Interaction Detection With Transformers (CVPR 2021 · similarity 0.94)
- Human-Object Interaction Detection via Disentangled Transformer (CVPR 2022 · similarity 0.92)
- MSTR: Multi-Scale Transformer for End-to-End Human-Object Interaction Detection (CVPR 2022 · similarity 0.92)
- Efficient Two-Stage Detection of Human-Object Interactions With a Novel Unary-Pairwise Transformer (CVPR 2022 · similarity 0.91)
- An Image-like Diffusion Method for Human-Object Interaction Detection (CVPR 2025 · 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.