Meshy T2：基於流匹配的快速原生網格生成

摘要

多邊形網格是現代 3D 管線中標準的表面表示方式，生成具有藝術家風格拓撲的高品質網格，對於電影、遊戲和互動式 3D 應用至關重要。主流方法將網格序列化為標記序列並以自迴歸方式解碼，這在推論時速度緩慢且對錯誤累積敏感，使其難以應用於互動式資產創作。我們提出 Meshy T2，一個基於流匹配的快速原生網格生成框架。其核心是一個頂點集網格 VAE，將網格編碼為每個頂點一個連續潛在標記，並在一次前向傳遞中解碼頂點、邊連通性和面繞序，在無需頂點量化或焊接的情況下，保留高精度幾何與藝術家創作的拓撲。生成過程以兩個流匹配模型的從粗到細級聯進行：首先，影像條件體素流將整體形狀草繪為粗略的佔用骨架；接著，網格流以每頂點潛在標記填充該骨架，並以影像、骨架及要求的頂點預算為條件。此設計提供了三項實用能力：透過基於流的平行合成實現互動式生成速度；透過要求的頂點預算實現有效的面數控制；以及對多部件資產的原生支援，其部件直接來自於生成的連通性。在我們的實驗中，Meshy T2 達到了最先進的幾何保真度，並在中位數 6 秒內完成端到端的影像到網格生成，比自迴歸基線快一個數量級以上。程式碼與權重將於 https://github.com/meshy-dev/meshy-t2 提供。

English

Polygonal meshes are the standard surface representation of modern 3D pipelines, and generating high-quality meshes with artist-style topology is essential for film, gaming, and interactive 3D applications. Mainstream approaches serialize a mesh into a token sequence and decode it autoregressively, which is slow at inference and sensitive to error accumulation, making them impractical for interactive asset creation. We present Meshy T2, a fast native mesh generation framework built on flow matching. At its core is a vertex-set mesh VAE that encodes a mesh into one continuous latent token per vertex and decodes vertices, edge connectivity, and face winding order in a single pass, preserving high-precision geometry and artist-authored topology without vertex quantization or welding. Generation proceeds as a coarse-to-fine cascade of two flow-matching models: an image-conditioned voxel flow first sketches the overall shape as a coarse occupancy scaffold, and a mesh flow then populates the scaffold with per-vertex latent tokens, conditioned on the image, the scaffold, and a requested vertex budget. This design delivers three practical capabilities: interactive generation speed through parallel flow-based synthesis; effective face-count control through the requested vertex budget; and native support for multi-part assets, whose components emerge directly from the generated connectivity. In our experiments, Meshy T2 achieves state-of-the-art geometric fidelity and completes end-to-end image-to-mesh generation within a median of 6 seconds, over an order of magnitude faster than autoregressive baselines. Code and weights will be available at https://github.com/meshy-dev/meshy-t2.