Research map / Video / Action recognition
Multi-Granularity Generator for Temporal Action Proposal
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 position embedding to video features (ablation) | MGG · ActivityNet-1.3 val | AUC: 65.59 -> 66.43 | Helped |
| Add bilinear matching between BaseNet convolutional representations (ablation) | MGG · ActivityNet-1.3 val | AUC: 65.88 -> 66.43 | Helped |
| Add the U-shape architecture in SPP (ablation) | MGG · ActivityNet-1.3 val | AUC: 65.02 -> 66.43 | Helped |
| Add FAP and TBA to SPP proposals (ablation) | MGG · ActivityNet-1.3 val | AUC: 64.31 -> 66.43 | Helped |
| Add SPP and TBA to FAP proposals (ablation) | MGG · ActivityNet-1.3 val | AUC: 59.91 -> 66.43 | Helped |
| Add the U-shape architecture (ablation) | MGG · ActivityNet-1.3 val, AN=100, tIoU=0.75, proposals 0-5s | recall: 0.15 -> 0.21 | Helped |
| Add TBA Stage 1 boundary adjustment (ablation) | MGG-F, end-to-end training · ActivityNet-1.3 val | AUC: 64.31 -> 65.54 | Helped |
| Add TBA Stage 2 middle-probability grouping adjustment (ablation) | MGG-F with Stage 1, end-to-end training · ActivityNet-1.3 val | AUC: 65.54 -> 66.43 | Helped |
| Train SPP and FAP jointly end-to-end rather than stagewise (ablation) | MGG with both TBA stages · ActivityNet-1.3 val | AUC: 66.28 -> 66.43 | Helped |
| Use coarse segment proposals alongside fine frame actionness (ablation) | MGG · ActivityNet-1.3 val | AUC: 64.31 -> 66.43 | Helped |
Closest papers by meaning
- Accurate Temporal Action Proposal Generation with Relation-Aware Pyramid Network (AAAI 2020 · similarity 0.88)
- Temporal Action Proposal Generation with Background Constraint (AAAI 2022 · similarity 0.87)
- Temporal Action Localization With Pyramid of Score Distribution Features (CVPR 2016 · similarity 0.87)
- Temporal Action Localization by Structured Maximal Sums (CVPR 2017 · similarity 0.87)
- Proposal-Based Multiple Instance Learning for Weakly-Supervised Temporal Action Localization (CVPR 2023 · similarity 0.86)
- Temporal Action Localization with Cross Layer Task Decoupling and Refinement (AAAI 2025 · similarity 0.86)
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.