Research map / Video / Action recognition
Minority-Oriented Vicinity Expansion with Attentive Aggregation for Video Long-Tailed Recognition
WonJun Moon, Hyun Seok Seong, Jae-Pil Heo
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 the two learnable feature aggregators (ablation) | VideoLT baseline · VideoLT | All accuracy: 0.516 -> 0.690 | Improved |
| Add the self-attentive aggregator (ablation) | VideoLT baseline · VideoLT | All accuracy: 0.516 -> 0.681 | Improved |
| Add the codebook-attentive aggregator (ablation) | VideoLT baseline · VideoLT | All accuracy: 0.516 -> 0.677 | Improved |
| Add both aggregators rather than either one alone (ablation) | VideoLT baseline · VideoLT | All accuracy: 0.681 / 0.677 -> 0.690 | Improved |
| Add dynamic extrapolation (ablation) | Both aggregators on VideoLT · VideoLT | Tail accuracy: 0.601 -> 0.632 | Improved |
| Add calibrated interpolation (ablation) | Both aggregators on VideoLT · VideoLT | Tail accuracy: 0.601 -> 0.632 | Improved |
| Combine dynamic extrapolation and calibrated interpolation (ablation) | Both aggregators on VideoLT · VideoLT | All accuracy: 0.690 -> 0.719 | Improved |
| Use prototypical self-attention instead of conventional MSA (ablation) | VideoLT with TSM features · VideoLT | All accuracy: 0.642 -> 0.669 | Improved |
| Use actual class sample counts rather than a linear tail-weight criterion (ablation) | MOVE · VideoLT | All accuracy: 0.699 -> 0.719 | Improved |
| Use MOVE instead of the strongest reported baseline | VideoLT with ResNet-101 features · VideoLT | Tail accuracy: 0.611 -> 0.644 | Improved |
| Use MOVE instead of Mixup | Imbalanced-MiniKinetics200 with ResNet-50, imbalance ratio 0.01 · Imbalanced-MiniKinetics200 | accuracy: 0.488 -> 0.509 | Improved |
| Use Mixup rather than MOVE | Imbalanced-MiniKinetics200 with ResNet-50, imbalance ratio 0.01 · Imbalanced-MiniKinetics200 | AP: 0.570 -> 0.570 | No clear effect |
Closest papers by meaning
- VideoLT: Large-Scale Long-Tailed Video Recognition (ICCV 2021 · similarity 0.91)
- SimLTD: Simple Supervised and Semi-Supervised Long-Tailed Object Detection (CVPR 2025 · similarity 0.88)
- Improving Visual Prompt Tuning by Gaussian Neighborhood Minimization for Long-Tailed Visual Recognition (NeurIPS 2024 · similarity 0.87)
- OFBD: Object-Focused Background Debiasing for Long-Tailed Learning (NeurIPS 2026 · similarity 0.86)
- Feature Fusion from Head to Tail for Long-Tailed Visual Recognition (AAAI 2024 · similarity 0.85)
- DiffuLT: Diffusion for Long-tail Recognition Without External Knowledge (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.
