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

Human pose and shape (HPS) estimation methods achieve remarkable results.However, current HPS benchmarks are mostly designed to test models in scenariosthat are similar to the training data. This can lead to critical situations inreal-world applications when the observed data differs significantly from thetraining data and hence is out-of-distribution (OOD). It is therefore importantto test and improve the OOD robustness of HPS methods. To address thisfundamental problem, we develop a simulator that can be controlled in afine-grained manner using interpretable parameters to explore the manifold ofimages of human pose, e.g. by varying poses, shapes, and clothes. We introducea learning-based testing method, termed PoseExaminer, that automaticallydiagnoses HPS algorithms by searching over the parameter space of human poseimages to find the failure modes. Our strategy for exploring thishigh-dimensional parameter space is a multi-agent reinforcement learningsystem, in which the agents collaborate to explore different parts of theparameter space. We show that our PoseExaminer discovers a variety oflimitations in current state-of-the-art models that are relevant in real-worldscenarios but are missed by current benchmarks. For example, it finds largeregions of realistic human poses that are not predicted correctly, as well asreduced performance for humans with skinny and corpulent body shapes. Inaddition, we show that fine-tuning HPS methods by exploiting the failure modesfound by PoseExaminer improve their robustness and even their performance onstandard benchmarks by a significant margin. The code are available forresearch purposes.

PoseExaminer: Automated Testing of Out-of-Distribution Robustness in Human Pose and Shape Estimation