Research map / Video / Action recognition
A Hierarchical Pose-Based Approach to Complex Action Understanding Using Dictionaries of Actionlets and Motion Poselets
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 garbage collector to the base model (ablation) | Base Model · sub-JHMDB | accuracy: 70.6% -> 72.7% | Helped |
| add multimodal actionlets to the base model (ablation) | Base Model · sub-JHMDB | accuracy: 70.6% -> 75.3% | Helped |
| add actionlets, garbage collector, and latent spatial assignments (ablation) | Base Model · sub-JHMDB | accuracy: 70.6% -> 77.5% | Helped |
| use the full descriptor rather than GEO descriptor only (ablation) | Base model with GC · sub-JHMDB | accuracy: 66.9% -> 72.7% | Helped |
| add garbage collector to spatio-temporal annotation (ablation) | Our model · Composable Activities, testing set | precision: 0.59 -> 0.62 | Helped |
| add garbage collector to spatio-temporal annotation (ablation) | Our model · Composable Activities, testing set | recall: 0.77 -> 0.78 | Helped |
| use self-paced initialization instead of random initialization (ablation) | latent v model · Composable Activities | accuracy: 46.3% -> 91.1% | Helped |
| use self-paced initialization instead of action-interval clustering (ablation) | latent v model · Composable Activities | accuracy: 54.8% -> 91.1% | Helped |
| remove spatial supervision and infer spatial assignments as latent variables (ablation) | Composable Activities model · Composable Activities | accuracy: 91.8% -> 91.1% | Hurt |
| use the proposed method rather than Lillo et al. without GC | action recognition model · Composable Activities | accuracy: 85.7% -> 91.1% | Helped |
| use the proposed method rather than the strongest reported baseline | action recognition model · Concurrent Actions | precision: 0.85 -> 0.92 | Helped |
| use the proposed method rather than the strongest reported baseline | action recognition model · sub-JHMDB | accuracy: 75.6% -> 77.5% | Helped |
Closest papers by meaning
- AVA: A Video Dataset of Spatio-Temporally Localized Atomic Visual Actions (CVPR 2018 · similarity 0.89)
- Action Motifs: Self-Supervised Hierarchical Representation of Human Body Movements (CVPR 2026 · similarity 0.89)
- Action Detail Matters: Refining Video Recognition with Local Action Queries (CVPR 2025 · similarity 0.86)
- Recovering Complete Actions for Cross-dataset Skeleton Action Recognition (NeurIPS 2024 · similarity 0.86)
- H-MoRe: Learning Human-centric Motion Representation for Action Analysis (CVPR 2025 · similarity 0.85)
- Skeleton-based Action Recognition with Non-linear Dependency Modeling and Hilbert-Schmidt Independence Criterion (AAAI 2025 · 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.