Research map / Video / Action recognition
Few-Shot Video Classification via Temporal Alignment
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 |
|---|---|---|---|
| Use adaptive temporal alignment with flexible start and end points (ablation) | TAM · Kinetics 1-shot | accuracy: 69.2 -> 73.0 | Helped |
| Use adaptive temporal alignment with flexible start and end points (ablation) | TAM · Kinetics 5-shot | accuracy: 80.6 -> 85.8 | Helped |
| Use adaptive temporal alignment with flexible start and end points (ablation) | TAM · Something V2 1-shot | accuracy: 39.6 -> 42.8 | Helped |
| Use adaptive temporal alignment with flexible start and end points (ablation) | TAM · Something V2 5-shot | accuracy: 49.0 -> 52.3 | Helped |
| Replace fixed diagonal matching with adaptive alignment (ablation) | TAM · Kinetics 1-shot | accuracy: 66.2 -> 73.0 | Helped |
| Replace mean frame-pair distance with adaptive temporal alignment (ablation) | TAM · Kinetics 1-shot | accuracy: 67.8 -> 73.0 | Helped |
| Replace minimum frame-pair distance with adaptive temporal alignment (ablation) | TAM · Kinetics 1-shot | accuracy: 52.4 -> 73.0 | Helped |
| Replace fixed diagonal matching with adaptive alignment (ablation) | TAM · Something V2 1-shot | accuracy: 38.3 -> 42.8 | Helped |
| Replace mean frame-pair distance with adaptive temporal alignment (ablation) | TAM · Something V2 1-shot | accuracy: 35.2 -> 42.8 | Helped |
| Replace minimum frame-pair distance with adaptive temporal alignment (ablation) | TAM · Something V2 1-shot | accuracy: 29.7 -> 42.8 | Helped |
| Fine-tune the backbone on meta-training data and model temporal relations | Few-shot video classification · Kinetics and Something V2 | accuracy: TSN++: 64.5, 77.9, 33.6, 43.0; TRN++: 68.4, 82.0, 38.6, 48.9 | Helped |
| Vary the continuous-relaxation smoothing parameter λ (ablation) | TAM · Kinetics and Something V2 | accuracy | No clear effect |
Closest papers by meaning
- Temporal Alignment Networks for Long-Term Video (CVPR 2022 · similarity 0.88)
- Learning by Aligning Videos in Time (CVPR 2021 · similarity 0.88)
- Temporal Alignment-Free Video Matching for Few-shot Action Recognition (CVPR 2025 · similarity 0.87)
- Searching for Better Spatio-temporal Alignment in Few-Shot Action Recognition (NeurIPS 2022 · similarity 0.87)
- Frame Order Matters: A Temporal Sequence-Aware Model for Few-Shot Action Recognition (AAAI 2025 · similarity 0.87)
- DejaVid: Encoder-Agnostic Learned Temporal Matching for Video Classification (CVPR 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.