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

Motion retargeting is the long-standing problem in character animation thatconsists in transferring and adapting the motion of a source character toanother target character. A typical application is the creation of motionsequences from off-the-shelf motions by transferring them onto new characters.Motion retargeting is also promising to increase interoperability of existinganimation systems and motion databases, as they often differ in the structureof the skeleton(s) considered. Moreover, since the goal of motion retargetingis to abstract and transfer motion dynamics, effective solutions might provideexpressive and powerful human motion models in which operations such ascleaning or editing are easier. In this article, we present a novel neuralnetwork architecture for retargeting that extracts an abstract representationof human motion agnostic to skeleton topology and morphology. Based ontransformers, our model is able to encode and decode motion sequences withvariable morphology and topology -- extending the current scope of retargeting-- while supporting skeleton topologies not seen during the training phase.More specifically, our model is structured as an autoencoder, and encoding anddecoding are separately conditioned on skeleton templates to extract andcontrol morphology and topology. Beyond motion retargeting, our model has manyapplications since our abstract representation is a convenient space to embedmotion data from different sources. It may potentially be benefical to a numberof data-driven methods, allowing them to combine scarce specialised motiondatasets (e.g. with style or contact annotations) and larger general motiondatasets, for improved performance and generalisation ability. Moreover, weshow that our model can be useful for other applications beyond retargeting,including motion denoising and joint upsampling.

HuMoT: Human Motion Representation using Topology-Agnostic Transformers for Character Animation Retargeting