Research map / Video / Video-language understanding
NewsNet: A Novel Dataset for Hierarchical Temporal Segmentation
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 |
|---|---|---|---|
| Use visual, audio, and text input instead of visual input alone | Boundary-free (SCRL) temporal segmentation · NewsNet in-domain, Scene | F1 score: 76.8 -> 78.3 | Helped |
| Use visual, audio, and text input instead of visual input alone | Boundary-free (SCRL) temporal segmentation · NewsNet in-domain, Story | F1 score: 71.2 -> 75.4 | Helped |
| Use visual, audio, and text input instead of visual input alone | Boundary-free (SCRL) temporal segmentation · NewsNet in-domain, Topic | F1 score: 62.9 -> 73.2 | Helped |
| Train and test on different news-program domains rather than the in-domain split | Boundary-free (SCRL), visual + audio + text · NewsNet, Scene | F1 score: 78.3 -> 76.0 | Hurt |
| Train and test on different news-program domains rather than the in-domain split | Boundary-free (SCRL), visual + audio + text · NewsNet, Story | F1 score: 75.4 -> 72.9 | Hurt |
| Train and test on different news-program domains rather than the in-domain split | Boundary-free (SCRL), visual + audio + text · NewsNet, Topic | F1 score: 73.2 -> 72.2 | Hurt |
| Add Hierarchical Ranking Loss to Multi-Label hierarchical modeling (ablation) | Multi-Label temporal segmentation · NewsNet, Scene | F1 score: 77.4 / 76.8 -> 79.6 / 76.9 | Helped |
| Add Hierarchical Ranking Loss to Multi-Label hierarchical modeling (ablation) | Multi-Label temporal segmentation · NewsNet, Story | F1 score: 74.3 / 74.3 -> 74.4 / 73.5 | Mixed |
| Add Hierarchical Ranking Loss to Multi-Label hierarchical modeling (ablation) | Multi-Label temporal segmentation · NewsNet, Topic | F1 score: 74.3 / 72.6 -> 77.8 / 73.1 | Helped |
| Add Hierarchical Ranking Loss to Multi-Head hierarchical modeling (ablation) | Multi-Head temporal segmentation · NewsNet, Scene | F1 score: 79.8 / 76.8 -> 80.3 / 76.9 | Helped |
| Add Hierarchical Ranking Loss to Multi-Head hierarchical modeling (ablation) | Multi-Head temporal segmentation · NewsNet, Story | F1 score: 74.5 / 73.7 -> 76.3 / 74.6 | Helped |
| Add Hierarchical Ranking Loss to Multi-Head hierarchical modeling (ablation) | Multi-Head temporal segmentation · NewsNet, Topic | F1 score: 76.6 / 70.4 -> 76.5 / 73.2 | Mixed |
Closest papers by meaning
- FastVID: Dynamic Density Pruning for Fast Video Large Language Models (NeurIPS 2025 · similarity 0.87)
- FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation (CVPR 2024 · similarity 0.86)
- HourVideo: 1-Hour Video-Language Understanding (NeurIPS 2024 · similarity 0.86)
- Minimizing Labeled, Maximizing Unlabeled: An Image-Driven Approach for Video Instance Segmentation (CVPR 2025 · similarity 0.86)
- MovieBench: A Hierarchical Movie Level Dataset for Long Video Generation (CVPR 2025 · similarity 0.86)
- ViCaS: A Dataset for Combining Holistic and Pixel-level Video Understanding using Captions with Grounded Segmentation (CVPR 2025 · 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.