Search papers, labs, and topics across Lattice.
LATO.2 introduces a factorized flow matching framework for 3D mesh generation that separates vertex generation from connectivity modeling, addressing issues of vertex drift and surface integrity found in joint latent space approaches. By utilizing dedicated VAEs for each stage and anchoring them to a shared coarse voxel scaffold, the method achieves higher-resolution meshes and enables topology-adaptive editing. Experimental results demonstrate that LATO.2 outperforms existing topology-aware mesh generators in both geometric fidelity and connectivity quality.
Factorizing mesh generation into vertex and connectivity flows allows for unprecedented resolution and seamless topology editing, setting a new standard in 3D modeling.
Flow matching over carefully designed latent representations has recently emerged as a powerful paradigm for topology-aware mesh generation. Existing approaches, however, model vertices and connectivity jointly in a joint latent space, entangling continuous vertex geometry with discrete combinatorial structure; this complicates flow learning and manifests as drifting vertices and broken surfaces. We present LATO.2, a factorized flow matching framework that decomposes mesh generation into a vertex flow followed by a connectivity flow conditioned on the realized vertices, with both stages anchored to a shared coarse voxel scaffold. Dedicated VAEs underpin the two stages, recovering vertices at sub-voxel precision and embedding discrete connectivity into a continuous latent space. We demonstrate two advantages unique to this factorization: (i) part-wise generation, in which the scaffold is partitioned and each part synthesized at full latent capacity, yielding substantially higher-resolution meshes than a monolithic latent permits; and (ii) topology-adaptive editing, in which manipulating first-stage vertices induces the corresponding connectivity without re-optimization. Experiments show that LATO.2 surpasses state-of-the-art topology-aware mesh generators in geometric fidelity and connectivity quality.
