Score: 2

Geospatial Foundational Embedder: Top-1 Winning Solution on EarthVision Embed2Scale Challenge (CVPR 2025)

Published: September 3, 2025 | arXiv ID: 2509.06993v1

By: Zirui Xu , Raphael Tang , Mike Bianco and more

BigTech Affiliations: Microsoft

Potential Business Impact:

Makes maps understand land better for many uses.

Business Areas:
Image Recognition Data and Analytics, Software

EarthVision Embed2Scale challenge (CVPR 2025) aims to develop foundational geospatial models to embed SSL4EO-S12 hyperspectral geospatial data cubes into embedding vectors that faciliatetes various downstream tasks, e.g., classification, regression, etc. In this technical report, we introduce our proposed method for the Top-1 winning solution on the Embed2Scale Challenge.

Country of Origin
🇺🇸 United States

Page Count
4 pages

Category
Computer Science:
CV and Pattern Recognition