GLAMR (Left) recovers human meshes in consistent global coordinates from videos captured by dynamic cameras and infills missing poses (transparent) due to various occlusions (obstruction, missed detection, outside field of view), while standard human mesh recovery methods (Right) fail to do so. Abstract We present an approach for 3D global human mesh recovery from monocular…
