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

We introduce CharacterGAN, a generative model that can be trained on only afew samples (8 - 15) of a given character. Our model generates novel posesbased on keypoint locations, which can be modified in real time while providinginteractive feedback, allowing for intuitive reposing and animation. Since weonly have very limited training samples, one of the key challenges lies in howto address (dis)occlusions, e.g. when a hand moves behind or in front of abody. To address this, we introduce a novel layering approach which explicitlysplits the input keypoints into different layers which are processedindependently. These layers represent different parts of the character andprovide a strong implicit bias that helps to obtain realistic results even withstrong (dis)occlusions. To combine the features of individual layers we use anadaptive scaling approach conditioned on all keypoints. Finally, we introduce amask connectivity constraint to reduce distortion artifacts that occur withextreme out-of-distribution poses at test time. We show that our approachoutperforms recent baselines and creates realistic animations for diversecharacters. We also show that our model can handle discrete state changes, forexample a profile facing left or right, that the different layers do indeedlearn features specific for the respective keypoints in those layers, and thatour model scales to larger datasets when more data is available.

CharacterGAN: Few-Shot Keypoint Character Animation and Reposing