ThroughputBench: How fast can a deep learning model map the Earth?
Published:
This post is published on geospatialml.com, the blog I write with Isaac Corley.
Caleb Robinson
Principal Research Scientist at Microsoft's AI for Good Lab - Geospatial ML center
- Seattle, Washington
- Blog
- Google Scholar
- ORCID
- Github
- X (formerly Twitter)
You May Also Enjoy
Q&A from the slop trenches Permalink
Published:
Between the two of us, we reviewed 22 paper submissions this summer, spread across NeurIPS, WACV, and TerraBytes (a geospatial workshop at ECCV). Fifteen of the 22 (68%) contained entirely fabricated citations,…
Sentinel-2 Paint: Recreate Any Image from Real Satellite Imagery Permalink
Published:
What colors does the Earth actually come in? We built Sentinel-2 Paint to find out: a script that scours the Sentinel-2 archive for patches of every color the planet shows, and a browser app for playing with the…
Similar but Different: A Benchmark for Measuring Whether Models Actually Use Multispectral Bands Permalink
Published:
The broad availability of multispectral data is one thing that separates satellite-based machine learning from mainstream computer vision. Sentinel-2 L2A data has 12 surface-reflectance bands, which include visible…
Gaussian Splat-based Satellite Image Super Resolution Permalink
Published:
Multi-temporal super-resolution of a Sentinel-2 scene. From left: a single S2 natural-color observation (10m), a 0.8m aerial basemap for visual reference, a bicubic 10× upsample of the S2 input, our Gaussian-splat…