MorphAny3D: Unleashing the Power of Structured Latent in 3D Morphing

TL;DR

MorphAny3D introduces a training-free framework utilizing structured latent representations and novel attention mechanisms to produce high-quality, semantically consistent, and temporally smooth 3D morphing sequences across categories.

Contribution

It proposes MorphAny3D, a novel approach that leverages structured latent features and attention mechanisms for improved 3D morphing without training, addressing pose ambiguity and cross-category challenges.

Findings

Achieves state-of-the-art morphing quality across categories.

Supports decoupled morphing and 3D style transfer.

Demonstrates generalization to other SLAT-based models.

Abstract

3D morphing remains challenging due to the difficulty of generating semantically consistent and temporally smooth deformations, especially across categories. We present MorphAny3D, a training-free framework that leverages Structured Latent (SLAT) representations for high-quality 3D morphing. Our key insight is that intelligently blending source and target SLAT features within the attention mechanisms of 3D generators naturally produces plausible morphing sequences. To this end, we introduce Morphing Cross-Attention (MCA), which fuses source and target information for structural coherence, and Temporal-Fused Self-Attention (TFSA), which enhances temporal consistency by incorporating features from preceding frames. An orientation correction strategy further mitigates the pose ambiguity within the morphing steps. Extensive experiments show that our method generates state-of-the-art morphing sequences, even for challenging cross-category cases. MorphAny3D further supports advanced applications such as decoupled morphing and 3D style transfer, and can be generalized to other SLAT-based generative models. Project page: https://xiaokunsun.github.io/MorphAny3D.github.io/.