Research map / Video / Action recognition
Deep Sequential Context Networks for Action Prediction
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 |
|---|---|---|---|
| Transfer sequential context from full videos to partial-video features | DeepSCN · UCF101, 10% observed | recognition accuracy: 45.02% | Helped |
| Add self-reconstruction to learn robust features (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 75.61 -> 81.31 | Helped |
| Add information transfer from full observations (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 76.55 -> 81.31 | Helped |
| Add label-based feature regularization (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 76.45 -> 81.31 | Helped |
| Add monotonic reconstruction-error constraints (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 76.50 -> 81.31 | Helped |
| Add one learned feature layer rather than using raw features only (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 75.39 -> 78.96 | Helped |
| Increase learned feature depth from 1 to 3 layers (ablation) | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 78.96 -> 83.36 | Helped |
| Use one SVM without requiring the test progress level | DeepSCN · UCF101, average over 10 observation ratios | recognition accuracy: 81.31 -> 80.55 | No clear effect |
| Use one SVM without requiring the test progress level | DeepSCN · Sports-1M, average over 10 observation ratios | recognition accuracy: 72.49 -> 71.84 | No clear effect |
| Use one SVM without requiring the test progress level | DeepSCN · BIT-Interaction, average over 10 observation ratios | recognition accuracy: 90.63 -> 85.94 | Hurt |
Closest papers by meaning
- No Frame Left Behind: Full Video Action Recognition (CVPR 2021 · similarity 0.89)
- Action Knowledge Transfer for Action Prediction with Partial Videos (AAAI 2019 · similarity 0.89)
- Unsupervised Deep Learning of Mid-Level Video Representation for Action Recognition (AAAI 2018 · similarity 0.88)
- Deep Alternative Neural Network: Exploring Contexts as Early as Possible for Action Recognition (NeurIPS 2016 · similarity 0.88)
- Action Detail Matters: Refining Video Recognition with Local Action Queries (CVPR 2025 · similarity 0.87)
- FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation (CVPR 2024 · 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.