Research map / Video / Video segmentation and tracking
Integrating Affordances and Attention Models for Short-Term Object Interaction Anticipation
Lorenzo Mur-Labadia, Rubén Martínez-Cantín, J.J. Guerrero, Giovanni Maria Farinella, Antonino Furnari
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 learned environment affordances | STAformer++ · Ego4D-STA v1 val | All mAP: 4.31 -> 4.66 | Improved |
| add fixed environment affordances and interaction hotspots | STAformer · Ego4D-STA v1 val | All mAP: 3.53 -> 3.77 | Improved |
| replace Faster R-CNN head with DETR head (ablation) | DINOv2 image-only model · Ego4D-STA v1 val | Noun mAP: 17.48 -> 29.33 | Improved |
| frame-guided attention temporal pooling instead of temporal convolution (ablation) | STAformer · Ego4D-STA v1 val | All mAP: 2.94 -> 3.39 | Improved |
| compute temporal pooling per scale (ablation) | STAformer++ · Ego4D-STA v1 val | All mAP: 4.07 -> 4.31 | Improved |
| dual image-video attention instead of sum fusion (ablation) | STAformer · Ego4D-STA v1 val | All mAP: 3.39 -> 3.47 | Improved |
| use both-direction dual attention rather than one direction (ablation) | STAformer · Ego4D-STA v1 val | All mAP: 3.01 -> 3.53 | Improved |
| EgoVideo (fine-tuned) instead of TimeSformer video encoder (ablation) | STAformer++ · Ego4D-STA v1 val | All mAP: 2.89 -> 3.65 | Improved |
| fine-tune last blocks of Swin-T image encoder (ablation) | STAformer++ image-only · Ego4D-STA v1 val | All mAP: 2.71 -> 2.85 | Improved |
| center prior to re-weight scores (ablation) | STAformer · Ego4D-STA v1 val | Noun mAP: 21.71 -> 17.70 | Hurt |
| hands proximity prior to re-weight scores (ablation) | STAformer · Ego4D-STA v1 val | Noun mAP: 21.71 -> 16.35 | Hurt |
| add interaction hotspots (ablation) | STAformer++ · Ego4D-STA v1 val | Noun mAP and N+V mAP: 32.07 -> 32.84 N mAP and 15.00 -> 15.67 N+V mAP | Improved |
Closest papers by meaning
- AFF-ttention! Affordances and Attention models for Short-Term Object Interaction Anticipation (ECCV 2024 · similarity 0.98)
- Anticipating Object Interactions Via Aggregation and Distillation of Spatio-Temporal Knowledge From Vision Language Models (TPAMI 2026 · similarity 0.92)
- Summarize the Past to Predict the Future: Natural Language Descriptions of Context Boost Multimodal Object Interaction Anticipation (CVPR 2024 · similarity 0.89)
- Vision-Guided Action: Enhancing 3D Human Motion Prediction with Gaze-informed Affordance in 3D Scenes (CVPR 2025 · similarity 0.87)
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.87)
- Learning from Observer Gaze: Zero-Shot Attention Prediction Oriented by Human-Object Interaction Recognition (CVPR 2024 · similarity 0.86)
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.
