Research map / Video / Video-language understanding
Multi-Scale 2D Temporal Adjacency Networks for Moment Localization With Natural Language
Songyang Zhang, Houwen Peng, Jianlong Fu, Yijuan Lu, Jiebo Luo
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 |
|---|---|---|---|
| use sparse single-scale map instead of dense single-scale map (ablation) | 2D-TAN · TACoS (C3D) | Rank1@0.5: 23.09 -> 28.57 | Improved |
| use sparse multi-scale map instead of sparse single-scale map (ablation) | 2D-TAN · TACoS (C3D) | Rank1@0.5: 28.57 -> 35.29 | Improved |
| use sparse multi-scale map instead of sparse single-scale map (ablation) | 2D-TAN · Charades-STA (C3D) | Rank5@0.7: 54.09 -> 46.48 | Hurt |
| add gated convolution (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 35.29 -> 35.77 | Improved |
| replace stacked convolution with max pooling for moment features (ablation) | MS-2D-TAN · Charades-STA (C3D) | Rank1@0.7: 23.25 -> 25.11 | Improved |
| replace stacked convolution with max pooling for moment features (ablation) | MS-2D-TAN · ActivityNet Captions (C3D) | Rank1@0.7: 29.21 -> 25.58 | Hurt |
| increase number of scales K from 1 to 5 (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 25.49 -> 35.77 | Improved |
| increase number of scales K from 5 to 7 (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 35.77 -> 32.97 | Hurt |
| use kernel 9 instead of kernel 1 (neighbor context) (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 6.17 -> 33.07 | Improved |
| increase window size N from 128 to 512 (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 33.34 -> 35.77 | Improved |
| increase window size N from 256 to 512 (ablation) | MS-2D-TAN · TACoS (C3D) | Rank1@0.5: 35.34 -> 35.77 | Improved |
| add 2D-TAN convolution to dense map on complex queries (ablation) | 2D-TAN · TACoS complex queries | Rank1@0.5: 8.43 -> 13.25 | Improved |
Closest papers by meaning
- Learning 2D Temporal Adjacent Networks for Moment Localization with Natural Language (AAAI 2020 · similarity 0.97)
- Localizing Moments in Video With Natural Language (ICCV 2017 · similarity 0.92)
- Transferable Video Moment Localization by Moment-Guided Query Prompting (AAAI 2024 · similarity 0.91)
- Structured Multi-Level Interaction Network for Video Moment Localization via Language Query (CVPR 2021 · similarity 0.90)
- Multi-Stage Aggregated Transformer Network for Temporal Language Localization in Videos (CVPR 2021 · similarity 0.90)
- Boundary-Aware Temporal Dynamic Pseudo-Supervision Pairs Generation for Zero-Shot Natural Language Video Localization (AAAI 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.
