Research map / Video / Video-language understanding
MSR-VTT: A Large Video Description Dataset for Bridging Video and Language
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 |
|---|---|---|---|
| add C3D temporal features to VGG-19 frame features (ablation) | SA-LSTM · MSR-VTT-10K | BLEU@4: 37.6 -> 40.5 | Helped |
| add C3D temporal features to VGG-19 frame features (ablation) | SA-LSTM · MSR-VTT-10K | METEOR: 28.9 -> 29.9 | Helped |
| use soft attention instead of mean pooling (ablation) | LSTM with C3D+VGG-19 features · MSR-VTT-10K | BLEU@4: 40.1 -> 40.5 | Helped |
| use soft attention instead of mean pooling (ablation) | LSTM with C3D+VGG-19 features · MSR-VTT-10K | METEOR: 29.5 -> 29.9 | Helped |
| replace VGG-19 with VGG-16 frame features (ablation) | SA-LSTM with C3D features · MSR-VTT-10K | BLEU@4: 40.5 -> 40.3 | Hurt |
| replace VGG-19 with VGG-16 frame features (ablation) | SA-LSTM with C3D features · MSR-VTT-10K | METEOR: 29.9 -> 29.7 | Hurt |
| increase LSTM hidden size from 128 to 512 (ablation) | MP-LSTM with C3D features · MSR-VTT-10K | BLEU@4: 32.5 -> 39.9 | Helped |
| increase LSTM hidden size from 128 to 512 (ablation) | MP-LSTM with C3D features · MSR-VTT-10K | METEOR: 26.6 -> 29.3 | Helped |
| use temporal C3D features instead of VGG-19 frame features (ablation) | MP-LSTM · MSR-VTT-10K | BLEU@4: 37.3 -> 39.9 | Helped |
| use temporal C3D features instead of VGG-19 frame features (ablation) | MP-LSTM · MSR-VTT-10K | METEOR: 28.7 -> 29.3 | Helped |
| use soft attention rather than mean pooling (ablation) | LSTM with VGG-19 features, documentary category · MSR-VTT-10K documentary | METEOR: 34.7 -> 28.2 | Hurt |
| use soft attention rather than mean pooling (ablation) | LSTM with C3D+VGG-19 features, news category · MSR-VTT-10K news | METEOR: 31.2 -> 36.2 | Helped |
Closest papers by meaning
- Vript: A Video Is Worth Thousands of Words (NeurIPS 2024 · similarity 0.90)
- Distilling Vision-Language Models on Millions of Videos (CVPR 2024 · similarity 0.89)
- MotionBench: Benchmarking and Improving Fine-grained Video Motion Understanding for Vision Language Models (CVPR 2025 · similarity 0.89)
- MMBench-Video: A Long-Form Multi-Shot Benchmark for Holistic Video Understanding (NeurIPS 2024 · similarity 0.88)
- Video-MME: The First-Ever Comprehensive Evaluation Benchmark of Multi-modal LLMs in Video Analysis (CVPR 2025 · similarity 0.88)
- Video-XL: Extra-Long Vision Language Model for Hour-Scale Video Understanding (CVPR 2025 · similarity 0.88)
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.