仅用5张稀疏图像实现360度全景视角合成，效果超越现有方法。
MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views
- 将3D高斯溅射与视频扩散模型结合，直接在隐空间生成新视角。
- 在DL3DV-10K数据集上实现360°视角合成，视觉质量领先。
- 适用于少视角、全场景重建，适合虚拟现实与数字孪生应用。
我们提出MVSplat360，一种基于稀疏观测的前馈式360°新视角合成方法。该任务因输入视图重叠少、信息不足而高度病态，传统方法难以获得高质量结果。MVSplat360通过融合几何感知3D重建与时间一致性视频生成，将前馈3D高斯溅射（3DGS）模型重构的特征直接映射至预训练稳定视频扩散（SVD）模型的隐空间，作为姿态与视觉引导信号，驱动去噪过程生成逼真且三维一致的新视角。模型端到端可训练，支持仅需5张稀疏输入即可渲染任意视角。为评估性能，我们引入基于挑战性DL3DV-10K数据集的新基准，结果表明其在广角甚至360°新视角合成任务中显著优于现有最先进方法。在RealEstate10K数据集上的实验也验证了模型有效性。视频演示见项目主页：https://donydchen.github.io/mvsplat360。
原文摘要 · Abstract (English)
We introduce MVSplat360, a feed-forward approach for 360° novel view synthesis (NVS) of diverse real-world scenes, using only sparse observations. This setting is inherently ill-posed due to minimal overlap among input views and insufficient visual information provided, making it challenging for conventional methods to achieve high-quality results. Our MVSplat360 addresses this by effectively combining geometry-aware 3D reconstruction with temporally consistent video generation. Specifically, it refactors a feed-forward 3D Gaussian Splatting (3DGS) model to render features directly into the latent space of a pre-trained Stable Video Diffusion (SVD) model, where these features then act as pose and visual cues to guide the denoising process and produce photorealistic 3D-consistent views. Our model is end-to-end trainable and supports rendering arbitrary views with as few as 5 sparse input views. To evaluate MVSplat360's performance, we introduce a new benchmark using the challenging DL3DV-10K dataset, where MVSplat360 achieves superior visual quality compared to state-of-the-art methods on wide-sweeping or even 360° NVS tasks. Experiments on the existing benchmark RealEstate10K also confirm the effectiveness of our model. The video results are available on our project page: https://donydchen.github.io/mvsplat360.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。