Research map / Video / Action recognition
Online Human Action Detection Using Joint Classification-Regression Recurrent Neural Networks
Yanghao Li, Cuiling Lan, Junliang Xing, Wenjun Zeng, Chunfeng Yuan, Jiaying Liu
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 the regression subnetwork with joint training (ablation) | CA-RNN · OAD | F1-Score: 0.596 -> 0.653 | Improved |
| add the regression subnetwork with joint training (ablation) | CA-RNN · OAD | SL-Score: 0.378 -> 0.418 | Improved |
| add the regression subnetwork with joint training (ablation) | CA-RNN · OAD | EL-Score: 0.382 -> 0.443 | Improved |
| add the Soft Selector between FC2 and FC3 (ablation) | JCR-RNN · OAD | F1-Score: 0.616 -> 0.653 | Improved |
| add the Soft Selector between FC2 and FC3 (ablation) | JCR-RNN · OAD | SL-Score: 0.389 -> 0.418 | Improved |
| add the Soft Selector between FC2 and FC3 (ablation) | JCR-RNN · OAD | EL-Score: 0.413 -> 0.443 | Improved |
| use end-to-end frame-wise RNN instead of sliding window RNN | RNN-SW · OAD | F1-Score: 0.600 -> 0.653 | Improved |
| use LSTM instead of SVM classifier in sliding window detector | SVM-SW · OAD | F1-Score: 0.540 -> 0.600 | Improved |
| enlarge sliding window from 10 to 40 frames (ablation) | RNN-SW · OAD | F1-Score: 0.600 -> 0.476 | Hurt |
| enlarge sliding window from 10 to 20 frames (ablation) | RNN-SW · OAD | F1-Score: 0.600 -> 0.627 | Improved |
| add regression subnetwork with joint training (ablation) | CA-RNN · G3D FPS | F1-Score: 0.378 -> 0.523 | Mixed |
| add regression subnetwork with joint training (ablation) | CA-RNN · G3D Golf (action-based F1) | action-based F1: 50.00 -> 70.00 | Improved |
Closest papers by meaning
- Modeling Temporal Dynamics and Spatial Configurations of Actions Using Two-Stream Recurrent Neural Networks (CVPR 2017 · similarity 0.91)
- Skeleton-Based Online Action Prediction Using Scale Selection Network (TPAMI 2019 · similarity 0.91)
- Co-Occurrence Feature Learning for Skeleton Based Action Recognition Using Regularized Deep LSTM Networks (AAAI 2016 · similarity 0.91)
- An End-to-End Spatio-Temporal Attention Model for Human Action Recognition from Skeleton Data (AAAI 2017 · similarity 0.91)
- Regularizing Long Short Term Memory With 3D Human-Skeleton Sequences for Action Recognition (CVPR 2016 · similarity 0.91)
- PREDICT & CLUSTER: Unsupervised Skeleton Based Action Recognition (CVPR 2020 · similarity 0.90)
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.
