MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Yuedong, Zheng, Chuanxia, Xu, Haofei, Zhuang, Bohan, Vedaldi, Andrea, Cham, Tat-Jen, Cai, Jianfei
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929582231257088
author Chen, Yuedong
Zheng, Chuanxia
Xu, Haofei
Zhuang, Bohan
Vedaldi, Andrea
Cham, Tat-Jen
Cai, Jianfei
author_facet Chen, Yuedong
Zheng, Chuanxia
Xu, Haofei
Zhuang, Bohan
Vedaldi, Andrea
Cham, Tat-Jen
Cai, Jianfei
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2411_04924
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views
Chen, Yuedong
Zheng, Chuanxia
Xu, Haofei
Zhuang, Bohan
Vedaldi, Andrea
Cham, Tat-Jen
Cai, Jianfei
Computer Vision and Pattern Recognition
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.
title MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2411.04924