Wukong's 72 Transformations: High-fidelity Textured 3D Morphing via Flow Models

TL;DR

WUKONG is a training-free framework that uses flow models to generate high-fidelity textured 3D morphing between prompts, avoiding manual correspondence and providing smooth, detailed transitions.

Contribution

It introduces a novel flow-based transformer approach for 3D morphing that eliminates manual matching, and incorporates semantic consistency for high-quality texture preservation.

Findings

Outperforms state-of-the-art methods in 3D morphing quality

Supports both global texture transitions and identity-preserving morphing

Achieves superior results across diverse geometry and texture variations

Abstract

We present WUKONG, a novel training-free framework for high-fidelity textured 3D morphing that takes a pair of source and target prompts (image or text) as input. Unlike conventional methods -- which rely on manual correspondence matching and deformation trajectory estimation (limiting generalization and requiring costly preprocessing) -- WUKONG leverages the generative prior of flow-based transformers to produce high-fidelity 3D transitions with rich texture details. To ensure smooth shape transitions, we exploit the inherent continuity of flow-based generative processes and formulate morphing as an optimal transport barycenter problem. We further introduce a sequential initialization strategy to prevent abrupt geometric distortions and preserve identity coherence. For faithful texture preservation, we propose a similarity-guided semantic consistency mechanism that selectively retains high-frequency details and enables precise control over blending dynamics. This empowers WUKONG to support both global texture transitions and identity-preserving texture morphing, catering to diverse generation needs. Extensive quantitative and qualitative evaluations demonstrate that WUKONG significantly outperforms state-of-the-art methods, achieving superior results across diverse geometry and texture variations.