< Explain other AI papers

EarthCrafter: Scalable 3D Earth Generation via Dual-Sparse Latent Diffusion

Shang Liu, Chenjie Cao, Chaohui Yu, Wen Qian, Jing Wang, Fan Wang

2025-07-25

EarthCrafter: Scalable 3D Earth Generation via Dual-Sparse Latent
  Diffusion

Summary

This paper talks about EarthCrafter, a new system designed to create detailed 3D models of large parts of Earth by separating the process into structure and texture generation using advanced machine learning techniques.

What's the problem?

Creating realistic and detailed 3D models of large geographic areas is very hard because the Earth's surface is huge and complex, and current methods either lack detail or require too much computing power.

What's the solution?

The researchers collected a massive aerial 3D dataset called Aerial-Earth3D and built EarthCrafter to process the data efficiently by compressing detailed 3D shapes and textures into smaller, manageable representations. They then used specialized models to generate the shape (structure) and appearance (texture) separately, making the generation faster and more accurate.

Why it matters?

This matters because EarthCrafter allows for scalable and high-quality 3D Earth models that can be used for urban planning, virtual tourism, or environmental studies, making it easier to understand and interact with our planet digitally.

Abstract

EarthCrafter, a framework using sparse-decoupled latent diffusion, efficiently generates large-scale 3D Earth models by separating structural and textural generation, leveraging the Aerial-Earth3D dataset.