SmoCap: Unified Scale-Pose Canonicalization with Proxy-Mapped Trust-Region QP

2026-05-20

AI Summary

Existing staged approaches to human body shape scaling and pose estimation often suffer from anatomical inconsistencies due to decoupled optimization, particularly inducing shape-pose compensation errors in poorly observed directions. This work proposes SmoCap, a novel framework that achieves the first unified, coupled estimation of shape and pose. By operating within a sparse control subspace, SmoCap jointly optimizes shape and pose through a constrained trust-region quadratic program augmented with an analytical proxy mapping. The method incorporates the Jacobian of this proxy mapping and a leakage-resistant mechanism to effectively stabilize estimation in weakly observed regions and enforce coherent structural motion. Experiments demonstrate high accuracy and efficiency: knee flexion RMSE is 2.9°, anthropometric endpoint error is approximately 3%, median marker RMSE is around 20 mm, and each frame requires only 0.204–0.332 ms with 2–3 iterations.

Abstract

Objective: Stage-wise workflows that separate model scaling and inverse kinematics can induce morphology-posture compensation, resulting in anatomically inconsistent yet numerically acceptable solutions, especially in weakly observed directions. We present SmoCap, a leakage-resistant canonicalization framework that estimates morphology and posture jointly in each local trust-region quadratic program (QP) within a sparse control subspace. Methods: SmoCap solves a constrained trust-region QP with analytical proxy-mapped pose and scale Jacobians. The low dimensional proxy map stabilizes weakly observed directions and drives coordinated structures. An optional pre-solve provides warm starts in difficult configurations. The framework is evaluated using cohort fluoroscopy knee motion, anthropometric ground truth, and extreme yoga sequences. Results: SmoCap achieved 2.9 degree knee flexion RMSE against fluoroscopy, and a pooled anthropometric endpoint error around 3%. In the leakage audit against segment wise scaling, SmoCap also reduced marker RMSE, FE error, and anthropometric endpoint error. Proxy coupling preserved expressive and coordinated spine motion with marginal fitting error increase (+0.14 mm, +0.6%) against baseline models in yoga ablation. Median marker RMSE was around 20 mm, and median runtime was 0.204-0.332 ms/frame, achieved with consistently 2-3 iterations. Conclusion: SmoCap provides an externally validated unified coupling-aware scale-pose framework, making externally consistent motion canonicalization practical at dataset scale.

Problem

• morphology-posture compensation
• anatomically inconsistent solutions
• weakly observed directions
• scale-pose canonicalization
• motion canonicalization

Innovation

• scale-pose canonicalization
• trust-region QP
• proxy mapping
• joint morphology-posture estimation
• leakage-resistant optimization