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.

TL;DR Given X_{t-s} and X_{t} 3D keypoints, calculate residual SMPL parameters from t-s to t.

To run sequential inference using the trained model, navigate to the cloned repository and execute the following command:

For detailed installation and data preparation, as well as instructions for training and evaluation, please refer to the GitHub repository.

If you find this work useful in your research, please consider citing:

@misc{LearnableSMPLify,
      title={Learnable SMPLify: A Neural Solution for Optimization-Free Human Pose Inverse Kinematics},
      author={Yuchen, Yang and Linfeng, Dong and Wei, Wang and Zhihang, Zhong and Xiao, Sun},
      year={2025},
      eprint={2508.13562},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}