Research map / Video / Action recognition
Dynamic Spatio-Temporal Specialization Learning for Fine-Grained Action Recognition
Tianjiao Li, Lin Geng Foo, Qiuhong Ke, Hossein Rahmani, Anran Wang, Jinghua Wang, Jun Liu
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 DSTS module | TPN · SSv2 | top-1 accuracy: 64.7 -> 67.2 | Improved |
| add DSTS module | Swin-B · SSv2 | top-1 accuracy: 69.6 -> 71.8 | Improved |
| add DSTS module | TPN · Diving48 | top-1 accuracy: 86.2 -> 88.4 | Improved |
| add DSTS module | Swin-B · Diving48 | top-1 accuracy: 80.5 -> 83.0 | Improved |
| add spatio-temporal specialization (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 87.2 -> 88.4 | Improved |
| add learned gates instead of fixed split (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 87.3 -> 88.4 | Improved |
| add synapse mechanism (activate one neuron) instead of averaging all neurons (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 86.5 -> 88.4 | Improved |
| add upstream-downstream learning (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 87.4 -> 88.4 | Improved |
| increase specialized neurons from 5 to 10 (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 87.3 -> 88.4 | Improved |
| increase specialized neurons from 10 to 15 (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 88.4 -> 88.3 | No clear effect |
| increase DSTS layers from 1 to 3 (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 87.5 -> 88.4 | Improved |
| increase DSTS layers from 3 to 5 (ablation) | DSTS (TPN) · Diving48 | top-1 accuracy: 88.4 -> 88.2 | No clear effect |
Closest papers by meaning
- Segmental Spatiotemporal CNNs for Fine-Grained Action Segmentation (ECCV 2016 · similarity 0.88)
- Learning Motion in Feature Space: Locally-Consistent Deformable Convolution Networks for Fine-Grained Action Detection (ICCV 2019 · similarity 0.87)
- SeFAR: Semi-supervised Fine-grained Action Recognition with Temporal Perturbation and Learning Stabilization (AAAI 2025 · similarity 0.87)
- Fine-Grained Temporal Contrastive Learning for Weakly-Supervised Temporal Action Localization (CVPR 2022 · similarity 0.87)
- FinePseudo: Improving Pseudo-Labelling through Temporal-Alignablity for Semi-Supervised Fine-Grained Action Recognition (ECCV 2024 · similarity 0.86)
- Learning From Temporal Gradient for Semi-Supervised Action Recognition (CVPR 2022 · similarity 0.86)
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.
