CreativeFlow: A One-to-Many Analogical Relation Transfer Method for 3D Asset Generation

Xuechen Li; Shuai Zhang; Nanxuan Zhao; Qing Chen

cs.AIOct 4, 2026

CreativeFlow: A One-to-Many Analogical Relation Transfer Method for 3D Asset Generation

Authors: Xuechen Li, Shuai Zhang, Nanxuan Zhao, Qing Chen

Organizations: College of Design and Innovation, Tongji University Shanghai, China · Adobe Research San Jose, California, United States

Abstract

Inspired by cognitive science, we present CREATIVEFLOW, an analogical generation framework that explicitly models analogical divergent thinking to mitigate creative homogenization in text-to-3D pipelines. Our method derives a series of meaningful yet relationally similar source-target asset pairs, each featuring distinct geometric configurations. Expert evaluations demonstrate that our framework substantially enhances creative novelty and visual fascination. This workflow and its resulting assets establish a foundational dataset and benchmark for future relation-aware 3D model training.

Figures & tables

Figure 1. (A) Standard text-to-3D models that collapse diverse intents into homogenized outputs, (B) CreativeFlow generates structurally diverse 3D designs approaching the richness of human references (C) and the illustration of creative homogenization(D). (A) Standard text-to-3D models that collapse diverse intents into homogenized outputs, (B) CreativeFlow generates structurally diverse 3D designs approaching the richness of human references (C) and the illustration of creative homogenization(D).

Figure 2. Examples of relation-guided analogical transfer. CreativeFlow transfers explicit aesthetic, structural/material, and mixed relations from a source asset to generate diverse yet recognizable target designs.

Figure 3. Our pipeline modeling $A:A^{\prime}::B:B^{\prime}$ . We extract source attribute set ${A^{\prime}}$ from $A$ via Attribute Inference, retrieve analogical concept set $B$ via Dual-Path Expansion, and solve for target attribute set $B^{\prime}$ to reconstruct diverse 3D target assets.

Figure 4. More results. One-to-many qualitative results across stools, handbags, and frogs. For each source, different analogical relations produce diverse target designs, with each row showing the source and multi-view renderings of the generated target.

Explore similar work

DreamStyle3D: Efficient 3D Stylized Asset Generation via Dual-Attention Disentanglement
Jul 27, 2026Kai Wang, Ziheng Ouyang, Xuying Zhang +23D Assets Fidelity
VCIP, Nankai University Tianjin, China · Zhongguancun Academy Beijing, China · Joy Future Academy, JD Beijing, China +1
AssetGen: Deployable 3D Asset Generation at Interactive Speed
May 22, 2026Dilin Wang, Xiaoyu Xiang, Kihyuk Sohn +143D Assets Speeds
Reality Labs, Meta
DiDE:Direct Injection with Color-Texture DEcoupling for 3D Stylization
Oct 1, 2026Tao Wu, Alexandra Gomez-Villa, Senmao Li +33D Generative Models 3D Assets
Computer Vision Center, Universitat Autònoma de Barcelona, Spain · Mohamed bin Zayed University of Artificial Intelligence, United Arab Emirates · Jilin University, China +1

Jul 27, 2026, cs.CVJ/K move · Enter open · S save

DreamStyle3D: Efficient 3D Stylized Asset Generation via Dual-Attention Disentanglement

Kai Wang, Ziheng Ouyang, Xuying Zhang, Ming-Ming Cheng+1

VCIP, Nankai University Tianjin, China · Zhongguancun Academy Beijing, China · Joy Future Academy, JD Beijing, China+1

3D Assets Fidelity

Open paper AskPDF arXiv

With the growth of gaming, animation, and virtual reality industries, the demand for efficient generation of stylized 3D assets is rapidly increasing. However, existing approaches still struggle to jointly preserve style fidelity, geometric consistency, and generation efficiency, as most of them still rely on indirect 2D-to-3D stylization pipelines. This motivates a native 3D stylization framework that can explicitly disentangle style from geometry while remaining efficient. To this end, we propose DreamStyle3D, an efficient framework for stylized 3D asset generation built on a Decoupled Dual Cross-Attention mechanism. Our method explicitly separates geometric and stylistic features to enable efficient style injection while preserving structural consistency, and further adopts a lightweight training strategy to enhance style consistency and model generalization. In addition, we build an automated data pipeline and construct a dataset of about 15K content-style-stylized triplets for training and evaluation. Extensive experiments demonstrate that our DreamStyle3D can generate high-fidelity, geometrically consistent stylized 3D assets within 10 seconds, substantially improving efficiency while maintaining superior style quality and offering a new solution for 3D content creation. The project is available at https://github.com/NK-JittorCV/nk-3D/tree/main/models/DreamStyle3D.

Abstract

Explore similar work

DreamStyle3D: Efficient 3D Stylized Asset Generation via Dual-Attention Disentanglement

AssetGen: Deployable 3D Asset Generation at Interactive Speed

DiDE:Direct Injection with Color-Texture DEcoupling for 3D Stylization