Research map / Video / Video-language understanding
Jointly Localizing and Describing Events for Dense Video Captioning
Yehao Li, Ting Yao, Yingwei Pan, Hongyang Chao, Tao Mei
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 |
|---|---|---|---|
| Jointly train event localization and caption generation with descriptiveness regression | DVC · ActivityNet Captions validation, learnt proposals | METEOR: 5.69 -> 6.93 | Improved |
| Jointly train event localization and caption generation with descriptiveness regression | DVC · ActivityNet Captions validation, GT proposals | METEOR: 8.88 -> 10.33 | Improved |
| Use descriptiveness-driven temporal attention instead of mean pooling (ablation) | LSTM captioner · ActivityNet Captions validation, learnt proposals | METEOR: 5.53 -> 5.79 | Improved |
| Add high-level semantic attributes (ablation) | DVC-D · ActivityNet Captions validation, learnt proposals | METEOR: 5.79 -> 6.14 | Improved |
| Replace cross-entropy sentence loss with METEOR-reward reinforcement learning (ablation) | DVC-D-A · ActivityNet Captions validation, learnt proposals | METEOR: 6.14 -> 6.93 | Improved |
| Use P3D ResNet pool5 features instead of C3D features (ablation) | DVC · ActivityNet Captions official test | METEOR: 10.36 -> 12.96 | Improved |
| Add conv9 anchor layer to conv3–conv8 (ablation) | DVC temporal event proposal · ActivityNet Captions validation | AUC: 55.73 -> 57.38 | Improved |
| Add conv10 anchor layer to conv3–conv9 (ablation) | DVC temporal event proposal · ActivityNet Captions validation | AUC: 57.38 -> 58.75 | Improved |
| Add conv11 anchor layer to conv3–conv10 (ablation) | DVC temporal event proposal · ActivityNet Captions validation | AUC: 58.75 -> 59.91 | Improved |
| Add conv12 anchor layer to conv3–conv11 (ablation) | DVC temporal event proposal · ActivityNet Captions validation | AUC: 59.91 -> 60.07 | Improved |
| Add descriptiveness regression to event proposal generation | DVC temporal event proposal · ActivityNet Captions validation | AUC | Improved |
| Use ground-truth proposals rather than learned proposals (ablation) | DVC · ActivityNet Captions validation | METEOR: 6.93 -> 10.33 | Improved |
Closest papers by meaning
- Weakly Supervised Dense Event Captioning in Videos (NeurIPS 2018 · similarity 0.94)
- Event-Equalized Dense Video Captioning (CVPR 2025 · similarity 0.93)
- Implicit Location-Caption Alignment via Complementary Masking for Weakly-Supervised Dense Video Captioning (AAAI 2025 · similarity 0.91)
- Dense Video Object Captioning from Disjoint Supervision (ICLR 2025 · similarity 0.91)
- Exploring Temporal Event Cues for Dense Video Captioning in Cyclic Co-Learning (AAAI 2025 · similarity 0.91)
- Do You Remember? Dense Video Captioning with Cross-Modal Memory Retrieval (CVPR 2024 · similarity 0.90)
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.
