VideoMV: Consistent Multi-View Generation Based on Large Video Generative Model

TL;DR

VideoMV introduces a novel multi-view generation framework leveraging video generative models and a 3D-aware sampling strategy, achieving fast, consistent multi-view image synthesis with high quality and reduced training time.

Contribution

The paper proposes a dense multi-view generation model fine-tuned from video generative models and introduces a 3D-aware denoising sampling method to enhance multi-view consistency.

Findings

Generates 24 dense views with high consistency.

Converges significantly faster (4 GPU hours) than state-of-the-art methods.

Outperforms existing approaches in quantitative metrics and visual quality.

Abstract

Generating multi-view images based on text or single-image prompts is a critical capability for the creation of 3D content. Two fundamental questions on this topic are what data we use for training and how to ensure multi-view consistency. This paper introduces a novel framework that makes fundamental contributions to both questions. Unlike leveraging images from 2D diffusion models for training, we propose a dense consistent multi-view generation model that is fine-tuned from off-the-shelf video generative models. Images from video generative models are more suitable for multi-view generation because the underlying network architecture that generates them employs a temporal module to enforce frame consistency. Moreover, the video data sets used to train these models are abundant and diverse, leading to a reduced train-finetuning domain gap. To enhance multi-view consistency, we introduce a 3D-Aware Denoising Sampling, which first employs a feed-forward reconstruction module to get an explicit global 3D model, and then adopts a sampling strategy that effectively involves images rendered from the global 3D model into the denoising sampling loop to improve the multi-view consistency of the final images. As a by-product, this module also provides a fast way to create 3D assets represented by 3D Gaussians within a few seconds. Our approach can generate 24 dense views and converges much faster in training than state-of-the-art approaches (4 GPU hours versus many thousand GPU hours) with comparable visual quality and consistency. By further fine-tuning, our approach outperforms existing state-of-the-art methods in both quantitative metrics and visual effects. Our project page is aigc3d.github.io/VideoMV.