Research map / Video / Action recognition
CLOT: Closed Loop Optimal Transport for Unsupervised Action Segmentation
Elena Bueno-Benito, Mariella Dimiccoli
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 Sliced Wasserstein distance cost refinement (ablation) | CLOT · Desktop Assembly | F1: 62.3 -> 72.6 | Improved |
| add Sliced Wasserstein distance cost refinement (ablation) | CLOT · 50Salads (Eval) | F1: 52.7 -> 63.2 | Improved |
| add Sliced Wasserstein distance cost refinement (ablation) | CLOT · Breakfast | MoF: 59.8 -> 60.1 | No clear effect |
| add feature dispatching (ablation) | CLOT · Breakfast | MoF: 51.9 -> 60.1 | Improved |
| add feature dispatching (ablation) | CLOT · 50Salads (Mid) | F1: 31.2 -> 46.6 | Improved |
| add feature dispatching (ablation) | CLOT · Desktop Assembly | mIoU: 44.3 -> 48.1 | Improved |
| add parallel decoder (ablation) | CLOT · 50Salads (Mid) | MoF: 44.9 -> 50.6 | Improved |
| add parallel decoder (ablation) | CLOT · YTI | MoF: 51.5 -> 54.4 | Improved |
| add refinement stage via cross-attention (ablation) | CLOT · 50Salads (Eval) | F1: 52.6 -> 63.2 | Improved |
| add refinement stage via cross-attention (ablation) | CLOT · YTI | MoF: 51.9 -> 54.4 | Improved |
| use parallel instead of autoregressive decoder (ablation) | CLOT · Desktop Assembly (activity) | F1: 25.3 -> 72.6 | Improved |
| use parallel instead of autoregressive decoder (ablation) | CLOT · 50Salads (Mid) (activity) | MoF: 52.3 -> 50.6 | Hurt |
Later papers that compared with this method
1 later accepted paper reports CLOT as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Learning Probabilistic Embeddings for Unsupervised Action Segmentation (ECCV 2026 · better on 4 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning Probabilistic Embeddings for Unsupervised Action Segmentation (ECCV 2026 · similarity 0.89)
- Temporally Consistent Unbalanced Optimal Transport for Unsupervised Action Segmentation (CVPR 2024 · similarity 0.87)
- Unsupervised Action Segmentation via Fast Learning of Semantically Consistent Actoms (AAAI 2024 · similarity 0.86)
- FACT: Frame-Action Cross-Attention Temporal Modeling for Efficient Action Segmentation (CVPR 2024 · similarity 0.85)
- Efficient and Effective Weakly-Supervised Action Segmentation via Action-Transition-Aware Boundary Alignment (CVPR 2024 · similarity 0.85)
- ActFusion: a Unified Diffusion Model for Action Segmentation and Anticipation (NeurIPS 2024 · similarity 0.85)
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.
