Research map / Video / Video-language understanding
What Actions Are Needed for Understanding Human Actions in Videos?
Gunnar A. Sigurdsson, Olga Russakovsky, Abhinav Gupta
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 training data (3/10 to full) (ablation) | Two-Stream (RGB) · Charades classification | mAP (%): 11.6 -> 15.6 | Improved |
| train on only the central 33% of each activity (ablation) | Two-Stream (RGB) · Charades classification | mAP (%): 15.9 -> 16.7 | Improved |
| ignore ambiguous boundary regions at evaluation (ablation) | Two-Stream · Charades per-frame localization | mAP (%): 9.6 -> 10.9 | Improved |
| ignore boundary regions in human consensus (ablation) | human annotators · Charades temporal IoU | IoU (%): 72.5 -> 79.8 | Improved |
| retrain on cropped person image (ablation) | Two-Stream (RGB) · Charades classification | mAP (%): 15.6 -> 17.9 | Improved |
| remove the person from test images (set to mean value) (ablation) | Two-Stream · Charades classification | mAP | Hurt |
| average predictions over a window of 4% of the video | Two-Stream, IDT, ActionVLAD, TFields · Charades | mAP | Mixed |
| more activities per video (1-4 to 14+) | Two-Stream · Charades classification | mAP (%): 22.1 -> 16.6 | Hurt |
Closest papers by meaning
- About Time: Advances, Challenges, and Outlooks of Action Understanding (IJCV 2025 · similarity 0.92)
- Object Level Visual Reasoning in Videos (ECCV 2018 · similarity 0.90)
- Moments in Time Dataset: One Million Videos for Event Understanding (TPAMI 2019 · similarity 0.89)
- Action Search: Spotting Actions in Videos and Its Application to Temporal Action Localization (ECCV 2018 · similarity 0.88)
- Disentangled Concepts Speak Louder Than Words: Explainable Video Action Recognition (NeurIPS 2025 · similarity 0.87)
- Learning to Evolve Scenes: Reasoning about Human Activities with Scene Graphs (NeurIPS 2026 · 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.
