IM-3D: Iterative Multiview Diffusion and Reconstruction for High-Quality 3D Generation

TL;DR

IM-3D introduces an efficient multi-view text-to-3D generation method that leverages video-based generators and Gaussian splatting, significantly improving quality and reducing computational costs compared to existing approaches.

Contribution

The paper presents a novel 3D generation pipeline that enhances multi-view synthesis using video generators and a robust reconstruction algorithm, reducing evaluations and artifacts.

Findings

10-100x fewer evaluations of 2D generators

Higher quality 3D outputs with fewer artifacts

Increased efficiency and yield of usable 3D assets

Abstract

Most text-to-3D generators build upon off-the-shelf text-to-image models trained on billions of images. They use variants of Score Distillation Sampling (SDS), which is slow, somewhat unstable, and prone to artifacts. A mitigation is to fine-tune the 2D generator to be multi-view aware, which can help distillation or can be combined with reconstruction networks to output 3D objects directly. In this paper, we further explore the design space of text-to-3D models. We significantly improve multi-view generation by considering video instead of image generators. Combined with a 3D reconstruction algorithm which, by using Gaussian splatting, can optimize a robust image-based loss, we directly produce high-quality 3D outputs from the generated views. Our new method, IM-3D, reduces the number of evaluations of the 2D generator network 10-100x, resulting in a much more efficient pipeline, better quality, fewer geometric inconsistencies, and higher yield of usable 3D assets.