Research map / Video / Action recognition
Video Playback Rate Perception for Self-Supervised Spatio-Temporal Representation Learning
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 |
|---|---|---|---|
| Pre-train with PRP instead of VCOP | C3D · UCF101 action recognition | accuracy (%): 65.6 -> 69.1 | Helped |
| Pre-train with PRP instead of VCOP | C3D · HMDB51 action recognition | accuracy (%): 28.4 -> 34.5 | Helped |
| Pre-train with PRP instead of VCOP | R3D · UCF101 action recognition | accuracy (%): 64.9 -> 66.5 | Helped |
| Pre-train with PRP instead of VCOP | R3D · HMDB51 action recognition | accuracy (%): 29.5 -> 29.7 | No clear effect |
| Pre-train with PRP instead of VCOP | R(2+1)D · UCF101 action recognition | accuracy (%): 72.4 -> 72.1 | Hurt |
| Pre-train with PRP instead of VCOP | R(2+1)D · HMDB51 action recognition | accuracy (%): 30.9 -> 35.0 | Helped |
| Use sampling interval {1,2,4,8} instead of {1,2} (ablation) | DP · UCF101 split 1 action recognition | accuracy (%): 68.3 -> 69.9 | Helped |
| Expand sampling intervals to {1,2,4,8,16} (ablation) | DP · UCF101 split 1 action recognition | accuracy (%): 69.9 -> 67.9 | Hurt |
| Add motion attention (ablation) | GP, sampling interval {1,2,4,8}, reconstruction rate 1 · UCF101 split 1 action recognition | accuracy (%): 67.1 -> 68.1 | Helped |
| Increase reconstruction rate from 1 to 4 with motion attention (ablation) | GP, sampling interval {1,2,4,8} · UCF101 split 1 action recognition | accuracy (%): 68.1 -> 68.4 | Helped |
| Combine discriminative and generative perception (ablation) | PRP · UCF101 split 1 action recognition | accuracy (%): 69.9 -> 70.9 | Helped |
Closest papers by meaning
- Self-Supervised Spatio-Temporal Representation Learning for Videos by Predicting Motion and Appearance Statistics (CVPR 2019 · similarity 0.89)
- RSPNet: Relative Speed Perception for Unsupervised Video Representation Learning (AAAI 2021 · similarity 0.88)
- Video Language Model Pretraining with Spatio-temporal Masking (CVPR 2025 · similarity 0.84)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.84)
- State Space Prompting via Gathering and Spreading Spatio-Temporal Information for Video Understanding (NeurIPS 2025 · similarity 0.84)
- SMILE: Infusing Spatial and Motion Semantics in Masked Video Learning (CVPR 2025 · similarity 0.84)
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.