Learnable SMPLify: A Neural Solution for Optimization-Free Human Pose Inverse Kinematics

TL;DR

Learnable SMPLify introduces a neural network-based approach to human pose estimation that replaces iterative optimization with a single-pass regression, achieving much faster runtimes while maintaining accuracy.

Contribution

It presents a novel neural framework that replaces traditional optimization in SMPLify with a single-pass model, improving speed and generalization in 3D human pose estimation.

Findings

Achieves nearly 200x faster runtime than SMPLify.

Generalizes well to unseen datasets like 3DPW and RICH.

Operates as a model-agnostic plug-in for existing estimators.

Abstract

In 3D human pose and shape estimation, SMPLify remains a robust baseline that solves inverse kinematics (IK) through iterative optimization. However, its high computational cost limits its practicality. Recent advances across domains have shown that replacing iterative optimization with data-driven neural networks can achieve significant runtime improvements without sacrificing accuracy. Motivated by this trend, we propose Learnable SMPLify, a neural framework that replaces the iterative fitting process in SMPLify with a single-pass regression model. The design of our framework targets two core challenges in neural IK: data construction and generalization. To enable effective training, we propose a temporal sampling strategy that constructs initialization-target pairs from sequential frames. To improve generalization across diverse motions and unseen poses, we propose a human-centric normalization scheme and residual learning to narrow the solution space. Learnable SMPLify supports both sequential inference and plug-in post-processing to refine existing image-based estimators. Extensive experiments demonstrate that our method establishes itself as a practical and simple baseline: it achieves nearly 200x faster runtime compared to SMPLify, generalizes well to unseen 3DPW and RICH, and operates in a model-agnostic manner when used as a plug-in tool on LucidAction. The code is available at https://github.com/Charrrrrlie/Learnable-SMPLify.