DeepAI AI Chat
Log In Sign Up

Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild

by   Dominik Kulon, et al.
Imperial College London

We introduce a simple and effective network architecture for monocular 3D hand pose estimation consisting of an image encoder followed by a mesh convolutional decoder that is trained through a direct 3D hand mesh reconstruction loss. We train our network by gathering a large-scale dataset of hand action in YouTube videos and use it as a source of weak supervision. Our weakly-supervised mesh convolutions-based system largely outperforms state-of-the-art methods, even halving the errors on the in the wild benchmark. The dataset and additional resources are available at


page 1

page 8


DeepHandMesh: A Weakly-supervised Deep Encoder-Decoder Framework for High-fidelity Hand Mesh Modeling

Human hands play a central role in interacting with other people and obj...

HandVoxNet: Deep Voxel-Based Network for 3D Hand Shape and Pose Estimation from a Single Depth Map

3D hand shape and pose estimation from a single depth map is a new and c...

End-to-end Weakly-supervised Multiple 3D Hand Mesh Reconstruction from Single Image

In this paper, we consider the challenging task of simultaneously locati...

3D Clothed Human Reconstruction in the Wild

Although much progress has been made in 3D clothed human reconstruction,...

Parallel mesh reconstruction streams for pose estimation of interacting hands

We present a new multi-stream 3D mesh reconstruction network (MSMR-Net) ...

Three Recipes for Better 3D Pseudo-GTs of 3D Human Mesh Estimation in the Wild

Recovering 3D human mesh in the wild is greatly challenging as in-the-wi...

HandTailor: Towards High-Precision Monocular 3D Hand Recovery

3D hand pose estimation and shape recovery are challenging tasks in comp...

Code Repositories


MobileHand: Real-time 3D Hand Shape and Pose Estimation from Color Image

view repo


This is an implementation of Joint Bone Loss by myself. I try it on my own pose estimation task and achieve good performance. So I share with you guys, enjoy~

view repo