Atlas3D: Physically Constrained Self-Supporting Text-to-3D for Simulation and Fabrication

TL;DR

Atlas3D introduces a method to generate physically stable, self-supporting 3D models from text prompts, ensuring models are suitable for simulation, fabrication, and real-world interaction.

Contribution

It presents a novel differentiable simulation-based loss and regularization technique to produce physically stable 3D models, enhancing existing text-to-3D methods.

Findings

Models maintain stability in physics simulations

Generated 3D objects can stand independently for printing

Method improves reliability of 3D models for real-world use

Abstract

Existing diffusion-based text-to-3D generation methods primarily focus on producing visually realistic shapes and appearances, often neglecting the physical constraints necessary for downstream tasks. Generated models frequently fail to maintain balance when placed in physics-based simulations or 3D printed. This balance is crucial for satisfying user design intentions in interactive gaming, embodied AI, and robotics, where stable models are needed for reliable interaction. Additionally, stable models ensure that 3D-printed objects, such as figurines for home decoration, can stand on their own without requiring additional supports. To fill this gap, we introduce Atlas3D, an automatic and easy-to-implement method that enhances existing Score Distillation Sampling (SDS)-based text-to-3D tools. Atlas3D ensures the generation of self-supporting 3D models that adhere to physical laws of stability under gravity, contact, and friction. Our approach combines a novel differentiable simulation-based loss function with physically inspired regularization, serving as either a refinement or a post-processing module for existing frameworks. We verify Atlas3D's efficacy through extensive generation tasks and validate the resulting 3D models in both simulated and real-world environments.