MorphAny3D
Unleashing the Power of Structured Latent in 3D Morphing
TL;DR: A training-free 3D morphing method that leverages the structured latent (SLAT) representation to achieve smooth and plausible deformations between diverse object categories.
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.
The leftmost and rightmost assets are the source and target objects, respectively. Click the cards to view the extracted GLB files. 3D assets are loading slowly, please be patient.
Application | Disentangled 3D Morphing
Application | Dual-Target 3D Morphing
Application | 3D Style Transfer
Generalization Ability | Hi3DGen
Generalization Ability | Text-to-3D Trellis
(a) Overview of our method. MorphAny3D generates a smooth and high-quality morphing sequence between diverse object categories by leveraging the SLAT representation without any training. (b) Morphing Cross-Attention (MCA) fuses information from the source and target objects in the cross-attention layers to ensure the structural coherence and aesthetics of the deformation. (c) Temporal-Fused Self-Attention (TFSA) enhances temporal smoothness by incorporating SLAT features from the previous morphing frame into the self-attention mechanism, enabling smooth transitions over time. (d) An orientation correction strategy inspired by statistical orientation distribution patterns in Trellis-generated assets is proposed to resolve abrupt orientation shifts.
If you find our work useful, please consider citing:
@article{sun2026morphany3d,
title = {MorphAny3D: Unleashing the Power of Structured Latent in 3D Morphing},
author = {Sun, Xiaokun and Cai, Zeyu and and Tang, Hao and Tai, Ying and
Yang, Jian and Zhang, Zhenyu},
journal = {arXiv preprint arXiv:2601.00204},
year = {2026}
}
The website template is borrowed from TRELLIS.