Research map / 3D vision / Human pose estimation
LSTM Pose Machines
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 LSTM module to the recurrent pose machine (ablation) | RPM · Penn Action test | PCK@0.2 Mean: 97.0 -> 97.7 | Helped |
| Add the LSTM module to the recurrent pose machine (ablation) | RPM · sub-JHMDB | PCK@0.2 Mean: 92.2 -> 93.6 | Helped |
| Increase recurrent iterations from 1 to 2 (ablation) | LSTM PM · sub-JHMDB | PCK@0.2 Mean: 91.1 -> 93.2 | Helped |
| Increase recurrent iterations from 2 to 5 (ablation) | LSTM PM · sub-JHMDB | PCK@0.2 Mean: 93.2 -> 93.6 | Helped |
| Increase recurrent iterations from 5 to 10 (ablation) | LSTM PM · sub-JHMDB | PCK@0.2 Mean: 93.6 -> 93.5 | Hurt |
| Use the recurrent LSTM architecture instead of multi-stage CPM inference | video pose estimation · 100-frame video clip | inference time per frame: 48.4ms -> 25.6ms | Helped |
Closest papers by meaning
- Deep Dual Consecutive Network for Human Pose Estimation (CVPR 2021 · similarity 0.88)
- Convolutional Pose Machines (CVPR 2016 · similarity 0.88)
- Recurrent Residual Module for Fast Inference in Videos (CVPR 2018 · similarity 0.87)
- Attention Mechanism Exploits Temporal Contexts: Real-Time 3D Human Pose Reconstruction (CVPR 2020 · similarity 0.87)
- Regularizing Long Short Term Memory With 3D Human-Skeleton Sequences for Action Recognition (CVPR 2016 · similarity 0.87)
- Recurrent 3D Pose Sequence Machines (CVPR 2017 · 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.