Deep Two-Stream Video Inference for Human Body Pose and Shape Estimation

TL;DR

This paper introduces DTS-VIBE, a novel two-stream transformer-based framework that fuses RGB and optical flow to improve the stability and accuracy of 3D human pose and shape estimation from videos.

Contribution

It presents a multi-modality approach combining RGB and optical flow with a transformer network for enhanced 3D human reconstruction from videos.

Findings

Outperforms state-of-the-art methods on Human3.6 and 3DPW datasets.

Improves temporal consistency and accuracy in 3D pose estimation.

Utilizes optical flow to leverage motion information between frames.

Abstract

Several video-based 3D pose and shape estimation algorithms have been proposed to resolve the temporal inconsistency of single-image-based methods. However it still remains challenging to have stable and accurate reconstruction. In this paper, we propose a new framework Deep Two-Stream Video Inference for Human Body Pose and Shape Estimation (DTS-VIBE), to generate 3D human pose and mesh from RGB videos. We reformulate the task as a multi-modality problem that fuses RGB and optical flow for more reliable estimation. In order to fully utilize both sensory modalities (RGB or optical flow), we train a two-stream temporal network based on transformer to predict SMPL parameters. The supplementary modality, optical flow, helps to maintain temporal consistency by leveraging motion knowledge between two consecutive frames. The proposed algorithm is extensively evaluated on the Human3.6 and 3DPW datasets. The experimental results show that it outperforms other state-of-the-art methods by a significant margin.