Research map / Vision-language and multimodal / Cross-modal learning
T2VLAD: Global-Local Sequence Alignment for Text-Video Retrieval
Xiaohan Wang, Linchao Zhu, Yi Yang
Check 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 |
|---|---|---|---|
| Use global-local alignment instead of the strongest reported baseline | T2VLAD · MSRVTT 1k-A, Text → Video | R@1: 26.6 -> 29.5 | Improved |
| Use global-local alignment instead of the strongest reported baseline | T2VLAD · MSRVTT 1k-B, Text → Video | R@1: 20.3 -> 26.1 | Improved |
| Use global-local alignment instead of the strongest reported baseline | T2VLAD · MSRVTT Full, Text → Video | R@1: 10.0 -> 12.7 | Improved |
| Use global-local alignment instead of the strongest reported baseline | T2VLAD · ActivityNet Captions, Text → Video | R@1: 22.7 -> 23.7 | Improved |
| Use global-local alignment instead of the strongest reported baseline | T2VLAD · LSMDC, Video → Text | R@1: 12.1 -> 14.2 | Improved |
| Add global alignment to local alignment (ablation) | T2VLAD · MSRVTT 1k-A, Text → Video | R@1: 24.3 -> 29.5 | Improved |
| Add global alignment to local alignment (ablation) | T2VLAD · MSRVTT 1k-A, Video → Text | R@1: 26.6 -> 31.8 | Improved |
| Add local alignment to global alignment (ablation) | T2VLAD · MSRVTT 1k-A, Text → Video | R@1: 22.2 -> 29.5 | Improved |
| Add local alignment to global alignment (ablation) | T2VLAD · MSRVTT 1k-A, Video → Text | R@1: 24.0 -> 31.8 | Improved |
| Use shared text-video VLAD rather than max pooling for video features (ablation) | T2VLAD · MSRVTT 1k-A, Text → Video | R@1: 27.4 -> 29.5 | Improved |
| Use shared text-video VLAD rather than max pooling for video features (ablation) | T2VLAD · MSRVTT 1k-A, Video → Text | R@1: 27.5 -> 31.8 | Improved |
| Share VLAD centers across text and video (ablation) | T2VLAD · MSRVTT 1k-A, Text → Video | R@1: 28.6 -> 29.5 | Improved |
Later papers that compared with this method
1 later accepted paper reports T2VLAD as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Audio-Visual Mismatch-Aware Video Retrieval via Association and Adjustment (ECCV 2022 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiscoVLA: Discrepancy Reduction in Vision, Language, and Alignment for Parameter-Efficient Video-Text Retrieval (CVPR 2025 · similarity 0.89)
- Towards Balanced Alignment: Modal-Enhanced Semantic Modeling for Video Moment Retrieval (AAAI 2024 · similarity 0.88)
- TOPA: Extending Large Language Models for Video Understanding via Text-Only Pre-Alignment (NeurIPS 2024 · similarity 0.88)
- Mitigating Semantic Collapse in Partially Relevant Video Retrieval (NeurIPS 2025 · similarity 0.88)
- Text Proxy: Decomposing Retrieval from a 1-to-N Relationship into N 1-to-1 Relationships for Text-Video Retrieval (AAAI 2025 · similarity 0.88)
- Rebalancing Contrastive Alignment with Bottlenecked Semantic Increments in Text-Video Retrieval (NeurIPS 2025 · 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.
