FreeArt3D: Training-Free Articulated Object Generation using 3D Diffusion

Chuhao Chen1, Isabella Liu1, Xinyue Wei1, 2, Hao Su1, 2, Minghua Liu2
1University of California San Diego   2Hillbot Inc  

SIGGRAPH Asia 2025

arXiv Paper Code 🤗 Demo Dataset & Results

FreeArt3D is a training-free framework that generates articulated 3D objects from a few images by leveraging a pre-trained 3D diffusion model for static objects. It jointly optimizes geometry, texture, and kinematics, achieving high-fidelity results across diverse categories without large datasets or task-specific training, outperforming prior state-of-the-art methods.

Input FreeArt3D Output
â—€ Slide to articulate the object â–¶

Abstract

Articulated 3D objects are central to many applications in robotics, AR/VR, and animation. Recent approaches to modeling such objects either rely on optimization-based reconstruction pipelines that require dense-view supervision or on feed-forward generative models that produce coarse geometric approximations and often overlook surface texture. In contrast, open-world 3D generation of static objects has achieved remarkable success, especially with the advent of native 3D diffusion models such as Trellis. However, extending these methods to articulated objects by training native 3D diffusion models poses significant challenges. In this work, we present FreeArt3D, a training-free framework for articulated 3D object generation. Instead of training a new model on limited articulated data, FreeArt3D repurposes a pre-trained static 3D diffusion model (e.g., Trellis) as a powerful shape prior. It extends Score Distillation Sampling (SDS) into the 3D-to-4D domain by treating articulation as an additional generative dimension. Given a few images captured in different articulation states, FreeArt3D jointly optimizes the object’s geometry, texture, and articulation parameters—without requiring task-specific training or access to large-scale articulated datasets. Our method generates high-fidelity geometry and textures, accurately predicts underlying kinematic structures, and generalizes well across diverse object categories. Despite following a per-instance optimization paradigm, FreeArt3D completes in minutes and significantly outperforms prior state-of-the-art approaches in both quality and versatility.

Method Overview

FreeArt3D takes sparse-view images of an articulated object in different joint states, and jointly optimizes the body geometry, the movable-part geometry and the joint parameters (axis, pivot, etc.). It builds a merged coarse occupancy grid by transforming the two parts according to the current joint parameters and states, and optimizes this grid using guidance from a pretrained 3D diffusion model TRELLIS. The result is then fed into the pretrained second-stage diffusion and VAE models to generate fine-grained geometry and realistic textures.

Pipeline Animation
Pipeline Overview
Visualization of the SDS Optimization Process

Results

Result 1
Box
Result 1
Cabinet
Result 1
Table
Result 1
Oven
PartNet-Mobility Results
Result 1
Diverse Categories of Objects
Diverse Categories of Objects

BibTeX

@InProceedings{chen2025freeart3d,
  title = {FreeArt3D: Training-Free Articulated Object Generation using 3D Diffusion},
  author = {Chen, Chuhao and Liu, Isabella and Wei, Xinyue and Su, Hao and Liu, Minghua},
  booktitle = {SIGGRAPH Asia 2025 Conference Papers},
  year = {2025}
}