Research map / Video / Action recognition
Further Understanding Videos through Adverbs: A New Video Task
Read the paperCheck 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 |
|---|---|---|---|
| add STSM (spatial-temporal separated module) (ablation) | RGB-BAUN · VAAD BA recognition | mAP: 26.0 -> 28.1 | Helped |
| add IRN (image relation network) (ablation) | RGB-BAUN · VAAD BA recognition | mAP: 25.7 -> 28.1 | Helped |
| add expression features (ablation) | RGB-BAUN · VAAD BA recognition | mAP: 27.1 -> 28.1 | Helped |
| add STSM (ablation) | RGB-BAUN · VAAD action recognition | accuracy: 52.7 -> 53.2 | Helped |
| add IRN (ablation) | RGB-BAUN · VAAD action recognition | accuracy: 52.4 -> 53.2 | Helped |
| add expression features (ablation) | RGB-BAUN · VAAD action recognition | accuracy: 53.0 -> 53.2 | Helped |
| pre-train on action then fine-tune for BA | Two-Stream-BAUN · VAAD BA recognition | mAP: 33.8 -> 35.5 | Helped |
| multi-task training with action-adverb conditional distribution | Two-Stream-BAUN · VAAD BA recognition | mAP: 33.8 -> 37.9 | Helped |
| pre-train on BA then fine-tune for action | Two-Stream-BAUN · HMDB-51 action recognition | accuracy: 81.7 -> 83.7 | Helped |
| pre-train on BA then fine-tune for action | Two-Stream-BAUN · UCF-101 action recognition | accuracy: 98.1 -> 98.4 | Helped |
| use BAUN instead of I3D | RGB input · VAAD BA recognition | mAP: 24.7 -> 28.1 | Helped |
| use BAUN instead of SlowFast R50 | RGB input · VAAD BA recognition | mAP: 27.7 -> 28.1 | No clear effect |
Closest papers by meaning
- Disentangled Concepts Speak Louder Than Words: Explainable Video Action Recognition (NeurIPS 2025 · similarity 0.88)
- Language Model Guided Interpretable Video Action Reasoning (CVPR 2024 · similarity 0.87)
- ActionAtlas: A VideoQA Benchmark for Domain-specialized Action Recognition (NeurIPS 2024 · similarity 0.87)
- Storyboard-guided Alignment for Fine-grained Video Action Recognition (NeurIPS 2025 · similarity 0.87)
- Open-Vocabulary Video Relation Extraction (AAAI 2024 · similarity 0.87)
- Does Video-Text Pretraining Help Open-Vocabulary Online Action Detection? (NeurIPS 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.