Search papers, labs, and topics across Lattice.
This paper introduces PolyFlow, a novel Transformer-based flow-matching framework that enables efficient parallel mesh generation by embedding discrete mesh topology into a continuous per-vertex state space. By projecting mesh vertex positions and normals into continuous embeddings, the framework overcomes the limitations of traditional autoregressive models, which suffer from significant computational overhead. Extensive evaluations show that PolyFlow outperforms state-of-the-art autoregressive methods in both Chamfer and Hausdorff distances, demonstrating its effectiveness in generating high-quality meshes rapidly and with precise control over resolution.
PolyFlow achieves rapid, high-quality mesh generation while outperforming traditional autoregressive models, thanks to its innovative continuous topology embedding.
Autoregressive Transformers dominate high-quality mesh generation by producing artist-worthy topologies, yet their inherent sequential decoding induces substantial computational overhead, falling orders of magnitude slower than parallel generative models. On the other hand, while continuous diffusion and flow-matching methods support efficient parallel synthesis across a variety of domains, they cannot be directly applied to meshes: mesh connectivity is inherently discrete and incompatible with standard continuous noise injection and denoising operations. To resolve this fundamental incompatibility, we introduce a compact topology embedder that projects discrete mesh vertex positions and normals into continuous per-vertex embeddings, where the original discrete adjacency information can be faithfully recovered via spacetime distance thresholding. After pretraining and freezing this embedder, any raw mesh can be fully converted into a continuous per-vertex state space unifying position, normal, and implicit topological attributes. Built upon this novel continuous mesh representation, we present PolyFlow, a Transformer-based flow-matching framework that achieves fully parallel vertex state denoising conditioned on extracted point-cloud features. During inference, our model completes generation rapidly via an ODE solver, and supports explicit, precise control over output mesh resolution by directly specifying the target vertex count. Extensive evaluations on the Toys4K benchmark demonstrate that PolyFlow surpasses state-of-the-art autoregressive baselines in both Chamfer Distance and Hausdorff Distance.
