Problem
Research questions and friction points this paper is trying to address.
text-to-3D generation
creative homogenization
3D asset generation
analogical relation transfer
CreativeFlow: A One-to-Many Analogical Relation Transfer Method for 3D Asset Generation
🤖 AI Summary
This study addresses the issue of creative homogenization in text-to-3D generation by proposing an analogy-based generative framework inspired by cognitive science to simulate divergent thinking. By explicitly modeling one-to-many relational transfer, the method generates pairs of 3D assets exhibiting distinct geometric configurations. Furthermore, this work constructs the first relation-aware 3D training dataset and evaluation benchmark. Experimental results demonstrate that, driven by analogical reasoning, the proposed approach significantly enhances the creative novelty and visual appeal of the generated outputs. Expert evaluations further validate its superior performance compared to existing methods.
📝 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.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Analogical Relation Transfer
3D Asset Generation
Divergent Thinking
Text-to-3D
Creative Homogenization
🔎 Similar Papers
No similar papers found.
