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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Appearance of dressed humans undergoes a complex geometric transformationinduced not only by the static pose but also by its dynamics, i.e., thereexists a number of cloth geometric configurations given a pose depending on theway it has moved. Such appearance modeling conditioned on motion has beenlargely neglected in existing human rendering methods, resulting in renderingof physically implausible motion. A key challenge of learning the dynamics ofthe appearance lies in the requirement of a prohibitively large amount ofobservations. In this paper, we present a compact motion representation byenforcing equivariance -- a representation is expected to be transformed in theway that the pose is transformed. We model an equivariant encoder that cangenerate the generalizable representation from the spatial and temporalderivatives of the 3D body surface. This learned representation is decoded by acompositional multi-task decoder that renders high fidelity time-varyingappearance. Our experiments show that our method can generate a temporallycoherent video of dynamic humans for unseen body poses and novel views given asingle view video.

Learning Motion-Dependent Appearance for High-Fidelity Rendering of Dynamic Humans from a Single Camera