English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

3D pose transfer is a challenging generation task that aims to transfer thepose of a source geometry onto a target geometry with the target identitypreserved. Many prior methods require keypoint annotations to findcorrespondence between the source and target. Current pose transfer methodsallow end-to-end correspondence learning but require the desired final outputas ground truth for supervision. Unsupervised methods have been proposed forgraph convolutional models but they require ground truth correspondence betweenthe source and target inputs. We present a novel self-supervised framework for3D pose transfer which can be trained in unsupervised, semi-supervised, orfully supervised settings without any correspondence labels. We introduce twocontrastive learning constraints in the latent space: a mesh-level loss fordisentangling global patterns including pose and identity, and a point-levelloss for discriminating local semantics. We demonstrate quantitatively andqualitatively that our method achieves state-of-the-art results in supervised3D pose transfer, with comparable results in unsupervised and semi-supervisedsettings. Our method is also generalisable to unseen human and animal data withcomplex topologies.

MAPConNet: Self-supervised 3D Pose Transfer with Mesh and Point Contrastive Learning