Research map / Video / Action recognition
A Comprehensive Review of Few-Shot Action Recognition
Yuyang Wanyan, Xiaoshan Yang, Weiming Dong, Changsheng Xu
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 |
|---|---|---|---|
| use CLIP-ViT-B backbone instead of ResNet-50 (TSAM vs MoLo) | few-shot action recognition · Kinetics 1-shot | accuracy: 74.0 -> 96.2 | Improved |
| use CLIP-ViT-B backbone instead of ResNet-50 (CLIP-FSAR vs TRX) | category prototype learning · SSv2-full 1-shot | accuracy: 42.0 -> 61.9 | Improved |
| use CLIP-FSAR text-guided prototypes vs TRX aggregation | prototype methods · Kinetics 5-shot | accuracy: 85.9 -> 95.0 | Improved |
| use VideoMAE 3D backbone (SAFSAR) instead of ResNet-50 MoLo | FSAR · SSv2-full 1-shot | accuracy: 56.6 -> 74.7 | Improved |
| hyper-relation set matching HyRSM instead of OTAM temporal alignment | alignment methods · SSv2-full 1-shot | accuracy: 42.8 -> 54.3 | Improved |
| HyRSM vs OTAM on 5-shot SSv2-full | alignment methods · SSv2-full 5-shot | accuracy: 52.3 -> 69.0 | Improved |
| Manta ImageNet-VM backbone vs TSAM on SSv2-full 5-shot | FSAR · SSv2-full 5-shot | accuracy: 74.6 -> 89.1 | Improved |
| Task-Adapter vs TSAM on SSv2-full 1-shot | CLIP-ViT-B FSAR · SSv2-full 1-shot | accuracy: 65.8 -> 71.3 | Improved |
| STRM vs TRX 5-shot on SSv2-full | ResNet-50 FSAR · SSv2-full 5-shot | accuracy: 64.6 -> 68.1 | Improved |
| STRM vs TRX 1-shot on SSv2-full | ResNet-50 FSAR · SSv2-full 1-shot | accuracy: 42.0 -> 43.1 | No clear effect |
Closest papers by meaning
- Matching Compound Prototypes for Few-Shot Action Recognition (IJCV 2024 · similarity 0.90)
- Few-Shot Action Recognition via Intra- and Inter-Video Information Maximization (TPAMI 2026 · similarity 0.90)
- CDFSL-V: Cross-Domain Few-Shot Learning for Videos (ICCV 2023 · similarity 0.90)
- Compound Prototype Matching for Few-Shot Action Recognition (ECCV 2022 · similarity 0.90)
- Few-Shot Action Recognition with Hierarchical Matching and Contrastive Learning (ECCV 2022 · similarity 0.89)
- Temporal Alignment-Free Video Matching for Few-shot Action Recognition (CVPR 2025 · 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.
