Research map / Video / Action recognition
Video Cloze Procedure for Self-Supervised Spatio-Temporal Learning
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 |
|---|---|---|---|
| Pre-train with VCP instead of VCOP | C3D · UCF101 | action recognition accuracy: 65.6 -> 68.5 | Helped |
| Pre-train with VCP instead of VCOP | C3D · HMDB51 | action recognition accuracy: 28.4 -> 32.5 | Helped |
| Pre-train with VCP instead of VCOP | R3D · UCF101 | action recognition accuracy: 64.9 -> 66.0 | Helped |
| Pre-train with VCP instead of VCOP | R(2+1)D · UCF101 | action recognition accuracy: 72.4 -> 66.3 | Hurt |
| Pre-train with VCP instead of VCOP | R(2+1)D · HMDB51 | action recognition accuracy: 30.9 -> 32.2 | Helped |
| Use spatial rotation only rather than random initialization (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 64.3 | Helped |
| Use spatial permutation only rather than random initialization (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 63.4 | Helped |
| Combine spatial rotation and permutation (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 66.0 | Helped |
| Use temporal remote shuffling only rather than random initialization (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 67.8 | Helped |
| Use temporal adjacent shuffling only rather than random initialization (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 65.0 | Helped |
| Combine temporal remote and adjacent shuffling (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 68.0 | Helped |
| Combine spatial rotation, spatial permutation, and both temporal operations (ablation) | VCP with C3D · UCF101 (first split) | action recognition accuracy: 62.0 -> 69.7 | Helped |
Closest papers by meaning
- Spatiotemporal Contrastive Video Representation Learning (CVPR 2021 · similarity 0.86)
- Self-Supervised Spatio-Temporal Representation Learning for Videos by Predicting Motion and Appearance Statistics (CVPR 2019 · similarity 0.86)
- Spatial-Then-Temporal Self-Supervised Learning for Video Correspondence (CVPR 2023 · similarity 0.86)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.84)
- Moving Off-the-Grid: Scene-Grounded Video Representations (NeurIPS 2024 · similarity 0.84)
- SMILE: Infusing Spatial and Motion Semantics in Masked Video Learning (CVPR 2025 · similarity 0.83)
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.