Abstract
We consider the problem of estimating a parametric model of 3D human meshfrom a single image. While there has been substantial recent progress in thisarea with direct regression of model parameters, these methods only implicitlyexploit the human body kinematic structure, leading to sub-optimal use of themodel prior. In this work, we address this gap by proposing a new technique forregression of human parametric model that is explicitly informed by the knownhierarchical structure, including joint interdependencies of the model. Thisresults in a strong prior-informed design of the regressor architecture and anassociated hierarchical optimization that is flexible to be used in conjunctionwith the current standard frameworks for 3D human mesh recovery. We demonstratethese aspects by means of extensive experiments on standard benchmark datasets,showing how our proposed new design outperforms several existing and popularmethods, establishing new state-of-the-art results. By considering jointinterdependencies, our method is equipped to infer joints even under datacorruptions, which we demonstrate by conducting experiments under varyingdegrees of occlusion.