English

Temporal-Aware Refinement for Video-based Human Pose and Shape Recovery

Computer Vision and Pattern Recognition 2023-11-17 v1

Abstract

Though significant progress in human pose and shape recovery from monocular RGB images has been made in recent years, obtaining 3D human motion with high accuracy and temporal consistency from videos remains challenging. Existing video-based methods tend to reconstruct human motion from global image features, which lack detailed representation capability and limit the reconstruction accuracy. In this paper, we propose a Temporal-Aware Refining Network (TAR), to synchronously explore temporal-aware global and local image features for accurate pose and shape recovery. First, a global transformer encoder is introduced to obtain temporal global features from static feature sequences. Second, a bidirectional ConvGRU network takes the sequence of high-resolution feature maps as input, and outputs temporal local feature maps that maintain high resolution and capture the local motion of the human body. Finally, a recurrent refinement module iteratively updates estimated SMPL parameters by leveraging both global and local temporal information to achieve accurate and smooth results. Extensive experiments demonstrate that our TAR obtains more accurate results than previous state-of-the-art methods on popular benchmarks, i.e., 3DPW, MPI-INF-3DHP, and Human3.6M.

Keywords

Cite

@article{arxiv.2311.09543,
  title  = {Temporal-Aware Refinement for Video-based Human Pose and Shape Recovery},
  author = {Ming Chen and Yan Zhou and Weihua Jian and Pengfei Wan and Zhongyuan Wang},
  journal= {arXiv preprint arXiv:2311.09543},
  year   = {2023}
}

Comments

20 pages, 12 figures

R2 v1 2026-06-28T13:22:55.074Z