Research map / Video / Action recognition
STEP: Spatio-Temporal Progressive Learning for Video Action Detection
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 |
|---|---|---|---|
| Add progressive spatial refinement steps (ablation) | STEP · UCF101 split 1 | frame-mAP: 51.5 -> 60.7 | Helped |
| Increase progressive steps from 2 to 3 (ablation) | STEP · UCF101 split 1 | frame-mAP: 60.7 -> 62.6 | Helped |
| Increase progressive steps from 3 to 4 (ablation) | STEP · UCF101 split 1 | frame-mAP: 62.6 -> 62.7 | Helped |
| Adaptively extend proposals over time instead of using only 6 frames (ablation) | STEP · UCF101 split 1 | frame-mAP | Helped |
| Use a direct 30-frame input without progressive temporal extension (ablation) | STEP · UCF101 split 1 | frame-mAP | Hurt |
| Use location anticipation instead of linear extrapolation for temporal extension (ablation) | STEP · UCF101 split 1, longer sequences | frame-mAP | Helped |
| Fuse RGB and optical flow at the network input rather than late (ablation) | STEP · UCF101 split 1 | frame-mAP: 70.7 -> 74.3 | Helped |
| Combine early fusion with two-stream detections (ablation) | STEP · UCF101 split 1 | frame-mAP: 74.3 -> 75.0 | Helped |
| Add scene context (ablation) | STEP · UCF101 split 1 | frame-mAP: +2.5% | Helped |
| Use hard-aware sampling (ablation) | STEP · UCF101 split 1 | frame-mAP: +1.5% | Helped |
| Increase IoU thresholds over progressive steps (ablation) | STEP · UCF101 split 1 | frame-mAP: +1% | Helped |
Closest papers by meaning
- Improving Action Localization by Progressive Cross-Stream Cooperation (CVPR 2019 · similarity 0.89)
- Finding Action Tubes with a Sparse-to-Dense Framework (AAAI 2020 · similarity 0.89)
- Stable Mean Teacher for Semi-supervised Video Action Detection (AAAI 2025 · similarity 0.88)
- Scene-Aware Spatiotemporal Generalization: Towards Robust Temporal Action Detection Across Domains (AAAI 2026 · similarity 0.88)
- AVA: A Video Dataset of Spatio-Temporally Localized Atomic Visual Actions (CVPR 2018 · similarity 0.88)
- Unsupervised Action Segmentation via Fast Learning of Semantically Consistent Actoms (AAAI 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.