DreamHOI: Subject-Driven Generation of 3D Human-Object Interactions with Diffusion Priors

TL;DR

DreamHOI is a zero-shot method that synthesizes realistic 3D human-object interactions from text descriptions by combining diffusion models with a dual implicit-explicit mesh representation.

Contribution

It introduces a novel dual implicit-explicit representation and a gradient optimization technique for realistic 3D HOI generation without extensive datasets.

Findings

Effective zero-shot synthesis of HOIs from text

Realistic 3D interactions generated with high fidelity

Outperforms existing methods in quality and diversity

Abstract

We present DreamHOI, a novel method for zero-shot synthesis of human-object interactions (HOIs), enabling a 3D human model to realistically interact with any given object based on a textual description. This task is complicated by the varying categories and geometries of real-world objects and the scarcity of datasets encompassing diverse HOIs. To circumvent the need for extensive data, we leverage text-to-image diffusion models trained on billions of image-caption pairs. We optimize the articulation of a skinned human mesh using Score Distillation Sampling (SDS) gradients obtained from these models, which predict image-space edits. However, directly backpropagating image-space gradients into complex articulation parameters is ineffective due to the local nature of such gradients. To overcome this, we introduce a dual implicit-explicit representation of a skinned mesh, combining (implicit) neural radiance fields (NeRFs) with (explicit) skeleton-driven mesh articulation. During optimization, we transition between implicit and explicit forms, grounding the NeRF generation while refining the mesh articulation. We validate our approach through extensive experiments, demonstrating its effectiveness in generating realistic HOIs.