AI Paper Atlas

Research map / 3D vision / Human pose estimation

MixSynthFormer: A Transformer Encoder-like Structure with Mixed Synthetic Self-attention for Efficient Human Pose Estimation

Yuran Sun, Alan William Dougherty, Zhuoying Zhang, Yi King Choi, Chuan Wu

ICCV 2023 · 3D vision › Human pose estimation · approach: Temporal video pose estimation · 13 citations (OpenAlex)

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 changedWhereMeasuredEffect
Remove spatial attention from MixSynth Attention (ablation)MixSynthFormer · Sub-JHMDBPCK@0.1 (%): 92.9 -> 95.1Improved
Add SE layer to attention matrix generation (ablation)MixSynthFormer · Sub-JHMDBPCK@0.1 (%): 94.2 -> 95.1Improved
Add weighted-sum combination of spatial and temporal branches (ablation)MixSynthFormer · Sub-JHMDBPCK@0.1 (%): 94.7 -> 95.1Improved
Add temporal attention (ablation)MixSynthFormer · Sub-JHMDBPCK@0.1 (%): 94.4 -> 95.1Improved
Use synthetic attention instead of vanilla self-attention (ablation)Temporal attention · 3DPWMPJPE: 92.6 -> 92.1; params 0.14M -> 0.10MImproved
Apply reduction factor 4 in temporal attention synthesis (ablation)MixSynthFormer · Human3.6MMPJPE: 52.2 -> 51.4Improved
Apply reduction factor on both spatial and temporal dimensions (ablation)MixSynthFormer · Human3.6MMPJPE: 52.2 -> 51.8; params 0.26M -> 0.18MMixed
Use MixSynthFormer instead of DeciWatchSampling-based pose estimation · Human3.6MMPJPE: 52.8 -> 50.9Improved
Use MixSynthFormer instead of DeciWatch on AIST++Sampling-based pose estimation · AIST++Accel error: 7.1 -> 4.7Improved
Use MixSynthFormer for longer prediction horizonsMotion prediction · Human3.6M 400 msMPJPE: 80 ms: 10.1 -> 8.4; 400 ms: 38.3 -> 37.4Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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