AI Paper Atlas

Research map / 3D vision / Human pose estimation

LaxMotion: Rethinking Supervision Granularity for 3D Human Motion Generation

Sheng Liu, Yuanzhi Liang, Sidan DU

ECCV 2026 · 3D vision › Human pose estimation · approach: Physics and sensor-based motion capture · 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
train with relaxed 2D supervision instead of 3D coordinatesMoMask · HumanML3DQM Score: 5.850 -> 8.805Improved
train with relaxed 2D supervision instead of 3D coordinatesMoMask · HumanML3DFID: 0.045 -> 0.054Hurt
fuse relaxed-supervision features and finetune with 3D supervisionMoMask · HumanML3DFID: 0.045 -> 0.034Improved
train MDM with relaxed supervisionMDM · HumanML3DFID: 0.544 -> 0.475Improved
train MoMask with relaxed supervision on KIT-MLMoMask · KIT-MLQM Score: 2.504 -> 4.982Improved
add trajectory-preserving relaxed supervision vs 2D-supervised methodMotion-2-to-3 · HumanML3D global motionFID: 0.321 -> 0.054Improved
use relaxed rather than full supervision signal (relaxed input) (ablation)3D finetune LaxMotion · HumanML3DQM Score: 5.879 -> 13.715Improved
use relaxed rather than full input (ablation)3D finetune LaxMotion · HumanML3DQM Score: 11.200 -> 13.715Improved
add cross-view plausibility loss L_rec (ablation)LaxMotion (MoMask-Based) · HumanML3DFID: 2.588 -> 0.054Improved
add orientation regularization L_ori (ablation)LaxMotion (MoMask-Based) · HumanML3DFID: 0.124 -> 0.054Improved
add feature consistency regularization L_feat (ablation)LaxMotion (MoMask-Based) · HumanML3DFID: 0.088 -> 0.054Improved
add feature consistency regularization L_feat (ablation)LaxMotion (MoMask-Based) · HumanML3DMModality: 2.086 -> 2.046Hurt

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