AI Paper Atlas

Research map / 3D vision / Human pose estimation

ViPNAS: Efficient Video Pose Estimation via Neural Architecture Search

Lumin Xu, Yingda Guan, Sheng Jin, Wentao Liu, Chen Qian, Ping Luo, Wanli Ouyang, Xiaogang Wang

CVPR 2021 · 3D vision › Human pose estimation · approach: Temporal video pose estimation · 59 citations (OpenAlex)

Check 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 changedWhereMeasuredEffect
Search temporal-level architectures and feature fusion under the same computation budget (ablation)ViPNAS · PoseTrack2018 val, ResNet-50, 0.82 GFLOPsmAP: 80.3 -> 81.6Improved
Search temporal-level architectures and feature fusion under the same computation budget (ablation)ViPNAS · PoseTrack2018 val, MobileNet-V3, 0.37 GFLOPsmAP: 77.2 -> 78.2Improved
Add temporal feature fusion (ablation)T-ViPNAS · PoseTrack2018 valmAPImproved
Allow different frame-specialized architectures instead of sharing the same spatial and temporal architectures (ablation)T-ViPNAS · PoseTrack2018 valmAP: at least 0.5% mAPImproved
Vary propagation length from 2 to 4 frames (ablation)ViPNAS temporal search · PoseTrack2018 val, 1.0 GFLOPsmAP: 81.7; 81.7; 81.4No clear effect
Replace SBL with S-ViPNASMobileNet-V3 pose estimator · PoseTrack2018 valTotal AP: 78.1 -> 81.1Improved
Replace SBL with T-ViPNASMobileNet-V3 pose estimator · PoseTrack2018 valGFLOPs: 4.1 -> 0.37Improved
Search a compact ResNet-50 architectureS-ViPNAS · COCO2017 val2017AP: 70.4 -> 71.0Improved
Search a compact MobileNet-V3 architectureS-ViPNAS · COCO2017 val2017AP: 64.7 -> 67.8Improved
Search the HRNet-W32 architectureS-ViPNAS · COCO2017 val2017AP: 74.4 -> 74.7Improved
Search the HRNet-W32 architectureS-ViPNAS · COCO2017 val2017GFLOPs: 7.10 -> 5.64Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy