Research map / Video / Action recognition
Multi-Instance Multi-Label Action Recognition and Localization Based on Spatio-Temporal Pre-Trimming for Untrimmed Videos
Read the paperCheck 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 multi-instance multi-label classification module (ablation) | PreTrimNet · THUMOS14 localization | Avg mAP IoU 0.1:0.5 (%): 36.31 -> 40.95 | Helped |
| add spatio-temporal pre-trimming (ablation) | PreTrimNet · THUMOS14 localization | Avg mAP IoU 0.1:0.5 (%): 32.40 -> 40.95 | Helped |
| add pre-trimming to model without MIML (ablation) | PreTrimNet · THUMOS14 localization | Avg mAP IoU 0.1:0.5 (%): 29.26 -> 36.31 | Helped |
| add MIML to model without pre-trimming (ablation) | PreTrimNet · THUMOS14 localization | Avg mAP IoU 0.1:0.5 (%): 29.26 -> 32.40 | Helped |
| fuse RGB, flow and pose streams vs RGB only (ablation) | PreTrimNet · THUMOS14 recognition | accuracy (%): 81.1 -> 89.2 | Helped |
| fuse RGB, flow and pose streams vs RGB only (ablation) | PreTrimNet · ActivityNet1.3 recognition | accuracy (%): 85.1 -> 93.3 | Helped |
| replace TSRNet with PreTrimNet | weakly supervised recognition · THUMOS14 recognition | accuracy (%): 87.1 -> 89.2 | Helped |
| replace TSRNet with PreTrimNet | weakly supervised localization · THUMOS14 localization | mAP@IoU 0.5 (%): 18.6 -> 23.09 | Helped |
| use local-global T-CAM with global refinement instead of local only (ablation) | PreTrimNet · THUMOS14 localization | mAP | Helped |
Closest papers by meaning
- Learning Transferable Self-Attentive Representations for Action Recognition in Untrimmed Videos with Weak Supervision (AAAI 2019 · similarity 0.90)
- Proposal-Based Multiple Instance Learning for Weakly-Supervised Temporal Action Localization (CVPR 2023 · similarity 0.90)
- ASM-Loc: Action-Aware Segment Modeling for Weakly-Supervised Temporal Action Localization (CVPR 2022 · similarity 0.90)
- PointTAD: Multi-Label Temporal Action Detection with Learnable Query Points (NeurIPS 2022 · similarity 0.90)
- Modeling Multi-Label Action Dependencies for Temporal Action Localization (CVPR 2021 · similarity 0.89)
- Weakly-Supervised Temporal Action Localization by Inferring Salient Snippet-Feature (AAAI 2024 · 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.