Research map / Label-efficient and robust learning / Self-supervised and contrastive
View-Invariant Skeleton Action Representation Learning via Motion Retargeting
Di Yang, Yaohui Wang, Antitza Dantcheva, Lorenzo Garattoni, Gianpiero Francesca, François Brémond
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 self reconstruction loss (ablation) | ViA · Smarthome CV2 | accuracy: 61.7 -> 62.9 | Improved |
| add self reconstruction loss (ablation) | ViA · Mixamo CV | accuracy: 71.7 -> 76.5 | Improved |
| add cycle reconstruction loss (ablation) | ViA · Smarthome CV2 | accuracy: 62.9 -> 63.8 | Improved |
| add cycle reconstruction loss (ablation) | ViA · Mixamo CV | accuracy: 76.5 -> 82.5 | Improved |
| add triplet loss (ablation) | ViA · Smarthome CV2 | accuracy: 63.8 -> 65.0 | Improved |
| add triplet loss (ablation) | ViA · Mixamo CV | accuracy: 82.5 -> 85.8 | Improved |
| add velocity loss (ablation) | ViA · Smarthome CV2 | accuracy: 65.0 -> 65.4 | Improved |
| add velocity loss (ablation) | ViA · Mixamo CV | accuracy: 85.8 -> 87.2 | Improved |
| add learnable Dc in latent motion disentanglement (ablation) | ViA · Mixamo motion retargeting | MSE: 2.42 -> 0.86 | Improved |
| use dictionary size K=32 instead of K=2 (ablation) | ViA · Mixamo motion retargeting | MSE: 1.16 -> 0.86 | Improved |
| use dictionary size K=32 instead of K=64 (ablation) | ViA · Mixamo motion retargeting | MSE: 0.89 -> 0.86 | Improved |
| add self-supervised pre-training on Posetics | ViA · Smarthome CV1 (fine-tuned) | accuracy: 22.9 -> 35.6 | Improved |
Closest papers by meaning
- Self-Supervised Skeleton-Based Action Representation Learning: A Benchmark and Beyond (IJCV 2026 · similarity 0.92)
- Self-Supervised 3D Skeleton Action Representation Learning With Motion Consistency and Continuity (ICCV 2021 · similarity 0.91)
- Revealing Key Details to See Differences: A Novel Prototypical Perspective for Skeleton-based Action Recognition (CVPR 2025 · similarity 0.89)
- Stitch, Contrast, and Segment: Learning a Human Action Segmentation Model Using Trimmed Skeleton Videos (AAAI 2025 · similarity 0.88)
- Heterogeneous Skeleton-Based Action Representation Learning (CVPR 2025 · similarity 0.88)
- Recovering Complete Actions for Cross-dataset Skeleton Action Recognition (NeurIPS 2024 · similarity 0.88)
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.
