AI Paper Atlas

Research map / 3D vision / Human pose estimation

PIAvatar: Physically Interactive Avatars via Deformation Gradient Decoupling

Sang-Hun Han, Min-Gyu Park, Jisu Shin, Seunghyun Shin, JINHWI PARK, Hae-Gon Jeon

ECCV 2026 · 3D vision › Human pose estimation · approach: Physics and sensor-based motion capture · 1 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 decoupled kinematic velocity and skeleton tracking instead of basic MPMPIAvatar · Avatar pose tracking, Animatable Gaussians, 100 stepsMSE (m): 0.079 -> 0.019Improved
use decoupled kinematic velocity instead of basic MPMPIAvatar · Avatar pose tracking, Animatable Gaussians, 400 stepsMSE (m): 0.332 -> 0.046Improved
use decoupled kinematic velocity instead of basic MPMPIAvatar · Avatar pose tracking, SMPL-X, 100 stepsMSE (m): 0.089 -> 0.013Improved
use decoupled kinematic velocity instead of basic MPMPIAvatar · Avatar pose tracking, SMPL-X, 400 stepsMSE (m): 0.248 -> 0.039Improved
use decoupled kinematic velocity instead of basic MPMPIAvatar · Avatar pose tracking, Animatable Gaussians, 400 stepsAcc@0.01: 0.020 -> 0.534Improved
use decoupled kinematic velocity instead of basic MPMPIAvatar · Avatar pose tracking, SMPL-X, 400 stepsAcc@0.01: 0.036 -> 0.583Improved
add decoupling and Kabsch tracking overhead to basic MPMPIAvatar · Runtime, Single SMPL-Xtime per step (ms): 0.170 -> 0.492Hurt
add decoupling and Kabsch tracking overhead to basic MPMPIAvatar · Runtime, Four AGtime per step (ms): 11.015 -> 15.259Hurt
add shape-preservation term (ablation)PIAvatar · Belly-ball contact (qualitative)residual dentImproved
add multi-field contact instead of single-field MPM contact (ablation)PIAvatar · Belly-ball contact and leg self-contact (qualitative)stickiness and penetrationImproved
optimize density with VLM-reward Bayesian optimization toward heavy vs lightPIAvatar · Ball mass BO with Qwen2.5-VLconverged mass (kg): 8.8 (heavy) vs 0.9 (light)Improved
optimize Young's modulus with VLM-reward Bayesian optimization toward stiff vs softPIAvatar · Belly stiffness BO with Qwen2.5-VLconverged E (Pa): 3.38x10^6 (stiff) vs 3.26x10^2 (soft)Improved

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