English

Test-Time Personalization with a Transformer for Human Pose Estimation

Computer Vision and Pattern Recognition 2021-11-09 v3

Abstract

We propose to personalize a human pose estimator given a set of test images of a person without using any manual annotations. While there is a significant advancement in human pose estimation, it is still very challenging for a model to generalize to different unknown environments and unseen persons. Instead of using a fixed model for every test case, we adapt our pose estimator during test time to exploit person-specific information. We first train our model on diverse data with both a supervised and a self-supervised pose estimation objectives jointly. We use a Transformer model to build a transformation between the self-supervised keypoints and the supervised keypoints. During test time, we personalize and adapt our model by fine-tuning with the self-supervised objective. The pose is then improved by transforming the updated self-supervised keypoints. We experiment with multiple datasets and show significant improvements on pose estimations with our self-supervised personalization.

Keywords

Cite

@article{arxiv.2107.02133,
  title  = {Test-Time Personalization with a Transformer for Human Pose Estimation},
  author = {Yizhuo Li and Miao Hao and Zonglin Di and Nitesh B. Gundavarapu and Xiaolong Wang},
  journal= {arXiv preprint arXiv:2107.02133},
  year   = {2021}
}

Comments

Project page: http://liyz15.github.io/TTP/

R2 v1 2026-06-24T03:54:20.299Z