Research map / Video / Video-language understanding
Progressive Spatio-Temporal Prototype Matching for Text-Video Retrieval
Pandeng Li, Chen-Wei Xie, Liming Zhao, Hongtao Xie, Jiannan Ge, Yun Zheng, Deli Zhao, Yongdong Zhang
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 object-phrase prototype matching to event-sentence matching (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 45.6 -> 48.2 | Improved |
| aggregate patches into object prototypes before matching with phrases (vs raw patches) (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 46.0 -> 48.2 | Improved |
| add frame decoder for object-to-frame aggregation (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 46.4 -> 48.2 | Improved |
| add attention mask in frame decoder (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 47.8 -> 48.2 | Improved |
| add residual with global frame token (Eq. 5) (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 46.6 -> 48.2 | Improved |
| use prototype-based event matching instead of aligning mean frame tokens with text token (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 46.5 -> 48.2 | Improved |
| use 3 event prototypes instead of 1 (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 46.3 -> 48.2 | Improved |
| increase object/phrase prototypes to 24/32 (ablation) | ProST · MSRVTT-9k text-to-video | R@1: 48.2 -> 46.0 | Hurt |
| replace TS2-Net with ProST (CLIP ViT-B/32) | text-video retrieval · MSRVTT-9k text-to-video | R@1: 47.0 -> 48.2 | Improved |
| replace TS2-Net with ProST (CLIP ViT-B/32) | text-video retrieval · DiDeMo text-to-video | R@1: 41.8 -> 44.9 | Improved |
| replace TVMM (reproduced) with ProST | text-video retrieval · MSRVTT-9k text-to-video | R@1: 45.8 -> 48.2 | Improved |
| increase spatial matching factor beta beyond optimum (1.5) (ablation) | ProST · MSRVTT-9k | R@1 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- TS2-Net — TS2-Net: Token Shift and Selection Transformer for Text-Video Retrieval (ECCV 2022) · better on 7 of 7 shared results
- X-Pool — X-Pool: Cross-Modal Language-Video Attention for Text-Video Retrieval (CVPR 2022) · better on 3 of 3 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
- Stitching Segments and Sentences towards Generalization in Video-Text Pre-training (AAAI 2024 · similarity 0.86)
- Generative Video Diffusion for Unseen Novel Semantic Video Moment Retrieval (AAAI 2025 · similarity 0.86)
- Bridging the Semantic Granularity Gap Between Text and Frame Representations for Partially Relevant Video Retrieval (AAAI 2025 · similarity 0.85)
- Video-ColBERT: Contextualized Late Interaction for Text-to-Video Retrieval (CVPR 2025 · similarity 0.85)
- OST: Refining Text Knowledge with Optimal Spatio-Temporal Descriptor for General Video Recognition (CVPR 2024 · similarity 0.85)
- Towards Balanced Alignment: Modal-Enhanced Semantic Modeling for Video Moment Retrieval (AAAI 2024 · 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.
