InHabit: Leveraging Image Foundation Models for Scalable 3D Human Placement

TL;DR

InHabit is a scalable, automatic method that leverages image foundation models to generate large-scale, photorealistic 3D human-scene interaction datasets, enhancing 3D understanding tasks.

Contribution

We introduce InHabit, a novel render-generate-lift pipeline that creates extensive 3D human-scene interaction data from scene renderings using vision-language models.

Findings

Produced 78K samples across 800 scenes with detailed 3D geometry and human models.

Improved 3D human-scene reconstruction and contact estimation when augmenting training data.

Data was preferred in 78% of cases over the state of the art in a user study.

Abstract

Training embodied agents to understand 3D scenes as humans do requires large-scale data of people meaningfully interacting with diverse environments, yet such data is scarce. Real-world motion capture is costly and limited to controlled settings, while existing synthetic datasets rely on simple geometric heuristics that ignore rich scene context. In contrast, 2D foundation models trained on internet-scale data have implicitly acquired commonsense knowledge of human-environment interactions. To transfer this knowledge into 3D, we introduce InHabit, a fully automatic and scalable data generator for populating 3D scenes with interacting humans. InHabit follows a render-generate-lift principle: given a rendered 3D scene, a vision-language model proposes contextually meaningful actions, an image-editing model inserts a human, and an optimization procedure lifts the edited result into physically plausible SMPL-X bodies aligned with the scene geometry. Applied to Habitat-Matterport3D, InHabit produces the first large-scale photorealistic 3D human-scene interaction dataset, containing 78K samples across 800 building-scale scenes with complete 3D geometry, SMPL-X bodies, and RGB images. Augmenting standard training data with our samples improves RGB-based 3D human-scene reconstruction and contact estimation, and in a perceptual user study our data is preferred in 78% of cases over the state of the art.