human-object interactions (HOIs) are crucial for human-centric scene
understanding applications such as human-centric visual generation, AR/VR, and
robotics. Since existing methods mainly explore capturing HOIs, rendering HOI
remains less investigated. In this paper, we address this ch