MARVEL-40M+: Multi-Level Visual Elaboration for High-Fidelity Text-to-3D Content Creation

TL;DR

This paper introduces MARVEL-40M+, a large-scale dataset with multi-level annotations for 3D assets, and a two-stage text-to-3D pipeline that enhances high-fidelity content creation from text prompts.

Contribution

It presents a novel multi-stage annotation pipeline combining VLMs and LLMs, and develops MARVEL-FX3D, a fast text-to-3D generation method, advancing dataset quality and generation speed.

Findings

MARVEL-40M+ outperforms existing datasets in annotation quality and diversity.

The pipeline achieves 72.41% win rate by GPT-4 and 73.40% by humans.

3D textured meshes generated within 15 seconds.

Abstract

Generating high-fidelity 3D content from text prompts remains a significant challenge in computer vision due to the limited size, diversity, and annotation depth of the existing datasets. To address this, we introduce MARVEL-40M+, an extensive dataset with 40 million text annotations for over 8.9 million 3D assets aggregated from seven major 3D datasets. Our contribution is a novel multi-stage annotation pipeline that integrates open-source pretrained multi-view VLMs and LLMs to automatically produce multi-level descriptions, ranging from detailed (150-200 words) to concise semantic tags (10-20 words). This structure supports both fine-grained 3D reconstruction and rapid prototyping. Furthermore, we incorporate human metadata from source datasets into our annotation pipeline to add domain-specific information in our annotation and reduce VLM hallucinations. Additionally, we develop MARVEL-FX3D, a two-stage text-to-3D pipeline. We fine-tune Stable Diffusion with our annotations and use a pretrained image-to-3D network to generate 3D textured meshes within 15s. Extensive evaluations show that MARVEL-40M+ significantly outperforms existing datasets in annotation quality and linguistic diversity, achieving win rates of 72.41% by GPT-4 and 73.40% by human evaluators. Project page is available at https://sankalpsinha-cmos.github.io/MARVEL/.