Research map / Detection and segmentation / Open-set and few-shot detection
Opening the Vocabulary of Egocentric Actions
Dibyadip Chatterjee, Fadime Sener, Shugao Ma, Angela Yao
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 OAP pretraining to S3D | S3D · Something-Else | Top-1: 49.3 -> 53.7 | Improved |
| Add OAP pretraining to Mformer-B | Mformer-B · Something-Else | Top-1: 60.2 -> 61.8 | Improved |
| Add guiding augmentations (temporal gradients + object mixing) with L_out only (ablation) | OAP (S3D) · Something-Else | Top-1: 50.1 -> 52.1 | Improved |
| Add L_out loss over supervised training (ablation) | S3D · Something-Else | Top-1: 49.3 -> 50.1 | Improved |
| Add L_in with object mixing (full OAP) vs without object mixing (ablation) | OAP (S3D) · Something-Else | Top-1: 50.3 -> 53.7 | Improved |
| Increase tau_in from 0.1 to 10.0 (ablation) | OAP (S3D) · Something-Else | Top-1: 53.7 -> 52.3 | Hurt |
| Lower tau_in from 0.1 to 0.01 (ablation) | OAP (S3D) · Something-Else | Top-1: 53.7 -> 53.5 | No clear effect |
| Use HOI object crops union hands instead of full image (ablation) | AOP · EPIC100-OV | HM: 7.5 -> 10.8 | Improved |
| Use hands crop only instead of full image (ablation) | AOP · EPIC100-OV | HM: 7.5 -> 6.8 | Hurt |
| Fine-tune base object vocabulary (ablation) | AOP · EPIC100-OV | HM: 10.6 -> 19.2 | Improved |
| Add verb-conditioned context prompts (ablation) | AOP · EPIC100-OV | HM: 18.8 -> 22.7 | Improved |
| Add learnable (non-verb-conditioned) context prompts (ablation) | AOP · EPIC100-OV | HM: 19.2 -> 18.8 | No clear effect |
Closest papers by meaning
- Do Egocentric Video-Language Models Truly Understand Hand-Object Interactions? (ICLR 2025 · similarity 0.91)
- Interactive Prototype Learning for Egocentric Action Recognition (ICCV 2021 · similarity 0.90)
- Symbiotic Attention with Privileged Information for Egocentric Action Recognition (AAAI 2020 · similarity 0.90)
- Action Scene Graphs for Long-Form Understanding of Egocentric Videos (CVPR 2024 · similarity 0.88)
- Modeling Fine-Grained Hand-Object Dynamics for Egocentric Video Representation Learning (ICLR 2025 · similarity 0.88)
- Summarize the Past to Predict the Future: Natural Language Descriptions of Context Boost Multimodal Object Interaction Anticipation (CVPR 2024 · similarity 0.88)
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.
