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

This paper considers a new problem of adapting a pre-trained model of humanmesh reconstruction to out-of-domain streaming videos. However, most previousmethods based on the parametric SMPL model \cite{loper2015smpl} underperform innew domains with unexpected, domain-specific attributes, such as cameraparameters, lengths of bones, backgrounds, and occlusions. Our general idea isto dynamically fine-tune the source model on test video streams with additionaltemporal constraints, such that it can mitigate the domain gaps withoutover-fitting the 2D information of individual test frames. A subsequentchallenge is how to avoid conflicts between the 2D and temporal constraints. Wepropose to tackle this problem using a new training algorithm named BilevelOnline Adaptation (BOA), which divides the optimization process of overallmulti-objective into two steps of weight probe and weight update in a trainingiteration. We demonstrate that BOA leads to state-of-the-art results on twohuman mesh reconstruction benchmarks.

Bilevel Online Adaptation for Out-of-Domain Human Mesh Reconstruction