AI Paper Atlas

Research map / Video / Action recognition

DisentangledTMR: Privacy-Preserving Skeleton Motion Retargeting via Factorized Transformers

Thomas Carr, Depeng Xu, Shuhan Yuan, Aidong Lu

ECCV 2026 · Video › Action recognition · approach: Skeleton-based action recognition · 0 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
use dual encoders with factorized decoder instead of single-encoder PMRDisentangledTMR · NTU60 X-Viewpre-trained AR: 35.7 -> 75.8Improved
use DisentangledTMR instead of DMRDMR · NTU60 X-Viewpre-trained RI: 25.7 -> 18.1Improved
use DisentangledTMR instead of PMRPMR · NTU60 X-Viewreconstruction MSE: 0.050 -> 0.030Improved
use DisentangledTMR instead of DMR for bone-length consistencyDMR · NTU60 X-Viewbone length error: 0.029 -> 0.016Improved
add Stage 2 and Stage 3 training after Stage 1 (ablation)DisentangledTMR · NTU60 X-Viewpre-trained AR: 57.3 -> ~82Improved
add Stages 2 and 3 after Stage 1 (privacy) (ablation)DisentangledTMR · NTU60 X-Viewpre-trained RI: 15.3Hurt
raise partial-retargeting ratio beta from 0.20 to 0.30DisentangledTMR · NTU60 X-Viewpre-trained AR: 76.8 -> 42.8Hurt
raise beta from 0.00 to 0.20DisentangledTMR · NTU60 X-Viewpre-trained RI: 75.4 -> 17.3Improved
raise beta to 1.00 (full retargeting)DisentangledTMR · NTU60 X-Viewpre-trained AR: 2.0Hurt
raise beta from 0.00 to 0.20 on NTU120DisentangledTMR · NTU120pre-trained RI: 67.6 -> 39.4Improved
raise beta from 0.00 to 0.20 on ETRIDisentangledTMR · ETRIpre-trained RI: 45.6 -> 35.7Improved
use the output-level supervision suite (NTU60) vs earlier checkpointsDisentangledTMR · NTU60 vs NTU120/ETRIpre-trained AR: 76.8%->42.8%Mixed

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