MeshPose: Unifying DensePose and 3D Body Mesh reconstruction

TL;DR

MeshPose unifies DensePose and Human Mesh Reconstruction by introducing new loss functions and a joint training approach, achieving accurate 2D and 3D body mesh localization suitable for real-time applications.

Contribution

We propose MeshPose, the first end-to-end system that combines DensePose and HMR with weak supervision, improving accuracy and efficiency for real-time 3D body mesh reconstruction.

Findings

Achieves competitive DensePose accuracy with low-poly mesh.

End-to-end training improves localization and reconstruction.

Lightweight model suitable for real-time AR applications.

Abstract

DensePose provides a pixel-accurate association of images with 3D mesh coordinates, but does not provide a 3D mesh, while Human Mesh Reconstruction (HMR) systems have high 2D reprojection error, as measured by DensePose localization metrics. In this work we introduce MeshPose to jointly tackle DensePose and HMR. For this we first introduce new losses that allow us to use weak DensePose supervision to accurately localize in 2D a subset of the mesh vertices ('VertexPose'). We then lift these vertices to 3D, yielding a low-poly body mesh ('MeshPose'). Our system is trained in an end-to-end manner and is the first HMR method to attain competitive DensePose accuracy, while also being lightweight and amenable to efficient inference, making it suitable for real-time AR applications.