# FLAT： 前馈潜码三角形泼溅实现几何精确场景生成

- 来源：HuggingFace Daily Papers（社区热门论文）
- 发布时间：2026-06-23 08:00
- AIHOT 分数：55
- AIHOT 链接：https://aihot.virxact.com/items/cmqrqfztx0mqkslp5syccjd0t
- 原文链接：https://arxiv.org/abs/2606.24876

## AI 摘要

FLAT 从单张图像直接解码视频扩散潜码中的三角形泼溅表面基元，首次实现前馈传递下从压缩潜码到显式三角形面片的映射。针对平面基元方向敏感、梯度流动困难问题，引入射线中心旋转参数化回归三角形，并设计乘积窗函数改进可微分三角形渲染的梯度流。标准基准上 FLAT 在保持视觉质量的同时取得显著更高的几何精度。轻量级测试时优化可将三角形网格转换为不透明、支持实时渲染的游戏引擎就绪表示。在相同训练设置下系统对比了 3DGS、2DGS 与三角形泼溅的表示权衡。

## 正文

Generating explorable 3D scenes from a single image requires strong generative priors and accurate geometric representations suitable for downstream use. Current video diffusion models offer high-quality generation and implicitly encode multi-view geometric structure in latent space. However, existing feedforward latent scene decoders typically output volumetric 3D Gaussians that lack a well-defined surface, limiting their use in simulation or standard graphics pipelines. This motivates decoding surface-aligned primitives that are not only renderable but also closer to explicit geometric assets. We ask whether compressed video diffusion latents can be mapped directly to explicit surface primitives in a single pass. To this end, we introduce FLAT and, for the first time, show that triangle splats can be decoded directly from video diffusion latents. Compared with decoding 3D Gaussians, predicting flat primitives is notoriously more challenging due to high sensitivity to primitive orientations, oftentimes leading to poor gradient flow. FLAT solves with two key ingredients: a ray-centered rotation parameterization for triangle regression and a novel product window function that improves gradient flow during differentiable triangle rendering. On standard benchmarks, FLAT achieves significantly better geometric accuracy while maintaining competitive visual quality compared to state-of-the-art feedforward baselines. We further show that a lightweight test-time refinement step converts the predicted triangle soup into a fully opaque, game-engine-ready representation that supports real-time rendering. By evaluating 3DGS, 2DGS, and triangle splatting variants under an identical training setup, we provide the first systematic analysis of representation tradeoffs in feedforward scene generation. The project page is available at https://flat-splat.github.io