MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views
About
We introduce MVSplat360, a feed-forward approach for 360{\deg} 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{\deg} 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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Novel View Synthesis | RE10K | SSIM75.6 | 142 | |
| Novel View Synthesis | DL3DV | PSNR16.37 | 84 | |
| Novel View Synthesis | T&T small-viewpoint set (O) | PSNR15.84 | 44 | |
| Novel View Synthesis | RE10K Small | PSNR13.68 | 38 | |
| New View Synthesis | T&T | LPIPS0.38 | 33 | |
| New View Synthesis | LLFF (R) | SSIM0.817 | 32 | |
| Novel View Synthesis | DL3DV S | LPIPS0.487 | 25 | |
| Novel View Synthesis | DTU small-viewpoint set (R) | PSNR15.31 | 24 | |
| New View Synthesis | DTU (R) | SSIM41.9 | 24 | |
| Novel View Synthesis | LLFF small-viewpoint set (R) | PSNR20.97 | 24 |