Research map / Video / Action recognition
Few-Shot Action Recognition via Intra- and Inter-Video Information Maximization
Huabin Liu, Tieyuan Chen, Yuxi Li, Site Li, John See, Weiyao Lin
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 video sampler only (ablation) | ProtoNet · HMDB51 5-way 1-shot | accuracy: 54.2 -> 59.0 | Improved |
| add action alignment only (ablation) | ProtoNet · SSv2 5-way 1-shot | accuracy: 34.2 -> 38.6 | Improved |
| add both video sampler and action alignment (ablation) | ProtoNet · HMDB51 5-way 1-shot | accuracy: 59.7 -> 59.9 | Improved |
| add L_intra and L_inter losses (ablation) | VIM sampler+alignment · UCF101 5-way 1-shot | accuracy: 83.5 -> 86.2 | Improved |
| add temporal selector (ablation) | VIM sampler · SSv2 5-way 1-shot | accuracy: 35.7 -> 37.5 | Improved |
| add spatial amplifier (ablation) | VIM sampler · SSv2 5-way 1-shot | accuracy: 36.9 -> 37.5 | No clear effect |
| add task-adaptive learner (ablation) | VIM sampler (TS+SA) · HMDB51 5-way 1-shot | accuracy: 59.1 -> 59.6 | Improved |
| replace spatial amplifying with cropping (ablation) | VIM · UCF101 5-way 5-shot | accuracy: 96.1 -> 93.3 | Hurt |
| increase scanned frames M from 20 to 32 (ablation) | VIM · UCF101 5-way 1-shot | accuracy: 86.2 -> 83.9 | Hurt |
| increase sampled frames T from 4 to 12 (ablation) | VIM · SSv2 5-way 1-shot | accuracy: 37.8 -> 41.2 | Improved |
| increase scanning input size from 64x64 to 224x224 (ablation) | VIM · SSv2 5-way 1-shot | accuracy: 40.8 -> 41.1 | No clear effect |
| extend VIM to TRX | TRX · SSv2 5-way 5-shot | accuracy: 59.1 -> 58.0 | Hurt |
Closest papers by meaning
- A Comprehensive Review of Few-Shot Action Recognition (IJCV 2025 · similarity 0.90)
- Matching Compound Prototypes for Few-Shot Action Recognition (IJCV 2024 · similarity 0.89)
- MGSampler: An Explainable Sampling Strategy for Video Action Recognition (ICCV 2021 · similarity 0.88)
- AVA: A Video Dataset of Spatio-Temporally Localized Atomic Visual Actions (CVPR 2018 · similarity 0.88)
- Temporal Alignment-Free Video Matching for Few-shot Action Recognition (CVPR 2025 · similarity 0.88)
- Action Detail Matters: Refining Video Recognition with Local Action Queries (CVPR 2025 · similarity 0.87)
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.
