Research map / Video / Action recognition
VPN: Learning Video-Pose Embedding for Activities of Daily Living
Srijan Das, Saurav Sharma, Rui Dai, François Brémond, Monique Thonnat
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 pose-driven attention network (ablation) | I3D · NTU-120 CS1 | accuracy: 77.0 -> 85.4 | Improved |
| add pose-driven attention network (ablation) | I3D · Smarthome CS | accuracy: 53.4 -> 56.4 | Improved |
| add spatial embedding (ablation) | I3D + attention network · Smarthome CS | accuracy: 56.4 -> 60.8 | Improved |
| add spatial embedding (ablation) | I3D + attention network · NTU-120 CS1 | accuracy: 85.4 -> 86.3 | Improved |
| use GCN instead of LSTM pose backbone (no coupler) (ablation) | VPN · Smarthome CS | accuracy: 57.1 -> 60.1 | Improved |
| use GCN instead of LSTM pose backbone (no coupler) (ablation) | VPN · NTU-120 CS1 | accuracy: 84.7 -> 85.6 | Improved |
| add spatio-temporal coupler (ablation) | VPN with GCN · NTU-120 CS2 | accuracy: 86.8 -> 87.8 | Improved |
| add spatio-temporal coupler (ablation) | VPN with GCN · Smarthome CS | accuracy: 60.1 -> 60.8 | Improved |
| add spatio-temporal coupler (ablation) | VPN with LSTM · NTU-120 CS1 | accuracy: 84.7 -> 85.3 | Improved |
| use normalized Euclidean embedding loss instead of bi-directional KL (ablation) | VPN · Smarthome CS | accuracy: 57.2 -> 60.8 | Improved |
| use bi-directional KL instead of one-directional KL (ablation) | VPN · NTU-120 CS1 | accuracy: 85.5 -> 86.1 | Improved |
| add spatial embedding to spatial-attention-only VPN with GCN (ablation) | VPN · Smarthome CS | accuracy: 49.1 -> 58.4 | Improved |
Closest papers by meaning
- VPN++: Rethinking Video-Pose Embeddings for Understanding Activities of Daily Living (TPAMI 2021 · similarity 0.94)
- Toyota Smarthome: Real-World Activities of Daily Living (ICCV 2019 · similarity 0.87)
- RPAN: An End-To-End Recurrent Pose-Attention Network for Action Recognition in Videos (ICCV 2017 · similarity 0.87)
- Convolutional Two-Stream Network Fusion for Video Action Recognition (CVPR 2016 · similarity 0.86)
- Spatiotemporal Residual Networks for Video Action Recognition (NeurIPS 2016 · similarity 0.86)
- HARDVS: Revisiting Human Activity Recognition with Dynamic Vision Sensors (AAAI 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.
