[CVPR 2020 Award Nominee] Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild
CrossMind.ai logo

[CVPR 2020 Award Nominee] Weakly-Supervised Mesh-Convolutional Hand Reconstruction in the Wild

Sep 24, 2020
|
30 views
|
Details
Authors: Dominik Kulon, Riza Alp Güler, Iasonas Kokkinos, Michael M. Bronstein, Stefanos Zafeiriou Description: 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 https://arielai.com/mesh_hands. Full Paper: https://arxiv.org/abs/2004.01946

Comments
loading...