Vivid-ZOO: Multi-View Video Generation with Diffusion Model

TL;DR

Vivid-ZOO introduces a diffusion-based method for generating high-quality, multi-view videos from text by leveraging pre-trained models and a novel factorization approach to ensure multi-view consistency and temporal coherence.

Contribution

The paper presents a new diffusion pipeline that combines pre-trained multi-view image and 2D video models for text-to-multi-view-video generation, reducing training costs and addressing domain gaps.

Findings

Generates high-quality multi-view videos with vivid motions.

Achieves multi-view consistency and temporal coherence.

Demonstrates effectiveness across diverse text prompts.

Abstract

While diffusion models have shown impressive performance in 2D image/video generation, diffusion-based Text-to-Multi-view-Video (T2MVid) generation remains underexplored. The new challenges posed by T2MVid generation lie in the lack of massive captioned multi-view videos and the complexity of modeling such multi-dimensional distribution. To this end, we propose a novel diffusion-based pipeline that generates high-quality multi-view videos centered around a dynamic 3D object from text. Specifically, we factor the T2MVid problem into viewpoint-space and time components. Such factorization allows us to combine and reuse layers of advanced pre-trained multi-view image and 2D video diffusion models to ensure multi-view consistency as well as temporal coherence for the generated multi-view videos, largely reducing the training cost. We further introduce alignment modules to align the latent spaces of layers from the pre-trained multi-view and the 2D video diffusion models, addressing the reused layers' incompatibility that arises from the domain gap between 2D and multi-view data. In support of this and future research, we further contribute a captioned multi-view video dataset. Experimental results demonstrate that our method generates high-quality multi-view videos, exhibiting vivid motions, temporal coherence, and multi-view consistency, given a variety of text prompts.