Search papers, labs, and topics across Lattice.
This paper introduces ZipTok3D, a novel 3D tokenizer that enhances reconstruction fidelity from significantly shorter token sequences by organizing object geometry into informative global-token prefixes. By employing a nested dropout technique during training, the model prioritizes essential geometric information, enabling high-quality reconstruction with drastically reduced token budgets. The results demonstrate that ZipTok3D achieves comparable reconstruction quality to existing methods while using up to 32 times fewer tokens, showcasing its efficiency in 3D generation tasks.
ZipTok3D reconstructs complex 3D objects with up to 32 times fewer tokens than traditional methods, revolutionizing the efficiency of 3D generation.
Compact token sequences are essential for efficient 3D generation. However, existing 3D tokenizers typically organize latent representations either over spatial regions or as fixed-size sets of global tokens, both suffering sharp reconstruction degradation when compressed to extremely low token budgets. In this paper, we present ZipTok3D, a 3D tokenizer designed for high-fidelity reconstruction from extremely short token sequences. Its key idea is to organize object geometry into progressively informative global-token prefixes and unfold these compact representations through iterative decoding. Specifically, nested dropout randomly truncates the latent sequence after encoding during training and requires each retained prefix to reconstruct the complete object, thereby prioritizing essential geometric information in the leading tokens. The decoder then repeatedly applies a parameter-shared Transformer block to recover fine-grained geometry from each prefix without a separate generative sampling stage. With the same token dimension, ZipTok3D achieves reconstruction quality comparable to the 32-token COD-VAE baseline using only one token on ShapeNet and four on TRELLIS, yielding $32\times$ and $8\times$ shorter token sequences, respectively.
