H2Avatar: Expressive Whole-Body Avatars from Monocular Video via Hierarchical Geometry and Hybrid Rendering | Synapse