AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning

Colorado J Reed, Ritwik Gupta, Shufan Li, Sarah Brockman, Christopher Funk, Brian Clipp, Kurt Keutzer, Salvatore Candido, Matt Uyttendaele, Trevor Darrell

ICCV 2023 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Masked autoencoder pretraining · 357 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add GSD positional encoding to vanilla MAE (ablation)Vanilla MAE · RESISC-45 kNN at 50% GSDkNN accuracy: 72.8 -> 75.4Improved
add progressive Laplacian decoder to vanilla MAE (ablation)Vanilla MAE · RESISC-45 kNN at 50% GSDkNN accuracy: 72.8 -> 75.3Improved
combine GSDPE and Laplacian decoder (ablation)Vanilla MAE · RESISC-45 kNN at 50% GSDkNN accuracy: 72.8 -> 78.1Improved
reduce decoder layers from 8 to 3 (ablation)Scale-MAE · RESISC-45 kNN at 100% GSDkNN accuracy: 78.9 -> 80.7Improved
use 75% mask rate instead of 70% or 80% (ablation)Scale-MAE · RESISC-45 kNN at 100% GSDkNN accuracy: 79.3 / 79.9 -> 80.7Improved
reconstruct both low and high frequency separately instead of low only (ablation)Scale-MAE · RESISC-45 kNN at 50% GSDkNN accuracy: 72.9 -> 78.1Improved
reconstruct separate low/high targets instead of a combined image (ablation)Scale-MAE · RESISC-45 kNN at 100% GSDkNN accuracy: 80.2 -> 80.7Improved
Scale-MAE versus SatMAE on kNN across eight datasetsScale-MAE · 8 remote sensing datasets kNNaverage kNN accuracy: +5.6%Improved
Scale-MAE versus ConvMAE on EuroSAT kNNScale-MAE · EuroSAT kNNaccuracy (%): 88.8 -> 86.7Hurt
Scale-MAE versus others on SpaceNet v1 at coarser GSDScale-MAE · SpaceNet v1 downscaledmIoU gap: +0.9 mIoU at 100%, +1.2 at 75%, +1.7 at 50%Improved
Scale-MAE decoder uses 3 layers instead of 8Scale-MAE · ViT-Large parametersparameters: 329.5M -> 322.9MImproved

Later papers that compared with this method

4 later accepted papers report ScaleMAE as a baseline in their results tables; 3 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy