Research map / Video / Video segmentation and tracking
Local Compressed Video Stream Learning for Generic Event Boundary Detection
Libo Zhang, Xin Gu, Congcong Li, Tiejian Luo, Heng Fan
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 local frames bag (k=2 versus k=0) (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.685 -> 0.741 | Improved |
| add LSTM temporal extractor instead of no temporal module (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.742 -> 0.768 | Improved |
| LSTM instead of Transformer (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.751 -> 0.768 | Improved |
| LSTM instead of GRU (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.760 -> 0.768 | No clear effect |
| LSTM instead of average pooling (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.699 -> 0.768 | Improved |
| add SCAM (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.653 -> 0.685 | Improved |
| add Gaussian-smoothed soft labels (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.760 -> 0.768 | Improved |
| larger backbone for motion vectors and residuals (ResNet-50/101 instead of ResNet-18) (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.768 -> 0.753 (ResNet-50), 0.750 (ResNet-101) | Hurt |
| bag size k=8 instead of k=10 or 12 (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.747 -> 0.768 (k=10), 0.745 -> 0.768 (k=12) | Improved |
| fuse I-frame, motion vector and residual features (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.650 -> 0.768 | Improved |
| sample 3 P-frames per GOP instead of 5 (ablation) | LCVSL · Kinetics-GEBD minval | F1: 0.761 -> 0.768 | Improved |
| use top-2 annotators' labels instead of top-1 for training (ablation) | LCVSL · Kinetics-GEBD minval | F1 | Improved |
Closest papers by meaning
- End-to-End Compressed Video Representation Learning for Generic Event Boundary Detection (CVPR 2022 · similarity 0.97)
- Fine-grained Dynamic Network for Generic Event Boundary Detection (ECCV 2024 · similarity 0.89)
- DiffDVC: Accurate Event Detection for Dense Video Captioning via Diffusion Models (AAAI 2025 · similarity 0.86)
- Boundary Denoising for Video Activity Localization (ICLR 2024 · similarity 0.85)
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.85)
- Temporal Action Detection Model Compression by Progressive Block Drop (CVPR 2025 · similarity 0.85)
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.
