Research map / Video / Action recognition
Frame-Level Label Refinement for Skeleton-Based Weakly-Supervised Action Recognition
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 |
|---|---|---|---|
| Use frame-level label refinement (ablation) | AGCN-based S-WTAL · BABEL Subset-1 | detection mAP Avg: 11.08 ± 1.08 -> 34.16 ± 1.26 | Helped |
| Use frame-level label refinement (ablation) | AGCN-based S-WTAL · BABEL Subset-2 | detection mAP Avg: 21.42 ± 2.89 -> 40.20 ± 3.09 | Helped |
| Use frame-level label refinement (ablation) | AGCN-based S-WTAL · BABEL Subset-3 | detection mAP Avg: 9.09 ± 1.59 -> 27.52 ± 1.23 | Helped |
| Use the proposed FLR method instead of FAC-Net | AGCN-based S-WTAL · BABEL Subset-1 | detection mAP Avg: 12.46 ± 0.63 -> 34.16 ± 1.26 | Helped |
| Use the proposed FLR method instead of CoLA | AGCN-based S-WTAL · BABEL Subset-2 | detection mAP Avg: 22.24 ± 2.41 -> 40.20 ± 3.09 | Helped |
| Use the proposed FLR method instead of FAC-Net | AGCN-based S-WTAL · BABEL Subset-3 | detection mAP Avg: 18.07 ± 4.25 -> 27.52 ± 1.23 | Helped |
| Remove joint optimization of frame-level pseudo labels (ablation) | FLR on BABEL · BABEL subsets, per-class detection mAP@0.1 | detection mAP@0.1 | Hurt |
| Remove multi-view learning loss (ablation) | FLR on BABEL · BABEL subsets, per-class detection mAP@0.1 | detection mAP@0.1 | Mixed |
| Set λpseudo to 0.2 or 0.5 instead of 1.0 (ablation) | FLR on BABEL · BABEL subsets, per-class detection mAP@0.1 | detection mAP@0.1 | Mixed |
| Set λKL to 0.2 or 1.0 instead of 0.5 (ablation) | FLR on BABEL · BABEL subsets, per-class detection mAP@0.1 | detection mAP@0.1 | Mixed |
| Set clean-frame selection α to 0.1 or 0.5 instead of 0.2 (ablation) | FLR on BABEL · BABEL subsets, per-class detection mAP@0.1 | detection mAP@0.1 | Mixed |
Closest papers by meaning
- Stitch, Contrast, and Segment: Learning a Human Action Segmentation Model Using Trimmed Skeleton Videos (AAAI 2025 · similarity 0.90)
- Towards To-a-T Spatio-Temporal Focus for Skeleton-Based Action Recognition (AAAI 2022 · similarity 0.90)
- Multi-Scale Spatial Temporal Graph Convolutional Network for Skeleton-Based Action Recognition (AAAI 2021 · similarity 0.89)
- InfoGCN: Representation Learning for Human Skeleton-Based Action Recognition (CVPR 2022 · similarity 0.89)
- Context Aware Graph Convolution for Skeleton-Based Action Recognition (CVPR 2020 · similarity 0.89)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.89)
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.