AI Paper Atlas

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

Learning Representations of Satellite Images From Metadata Supervision

Jules Bourcier, Gohar Dashyan, Karteek Alahari, Jocelyn Chanussot

ECCV 2024 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 13 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 SimCLR image-image loss to metadata-image contrastive pretraining (SatMIPS vs SatMIP)SatMIP · fMoW kNNmacro F1: 55.2 -> 62.3Improved
add metadata supervision to SimCLR (SatMIPS vs SimCLR)SimCLR · fMoW kNNmacro F1: 61.1 -> 62.3Improved
metadata-image contrast (SatMIP) instead of SimCLR image-only on fMoWSatMIP · fMoW kNNmacro F1: 61.1 -> 55.2Hurt
concatenate metadata features with image features at downstream linear probing (ablation)SatMIP · fMoW linear probemacro F1: 59.3 -> 63.1Improved
concatenate metadata features with image features on EuroSAT (ablation)SatMIPS · EuroSAT linear probeaccuracy: 95.8 -> 96.4Improved
add hierarchical pretraining from SLIP-YFCCSatMIP · fMoW kNNmacro F1: 55.2 -> 61.2Improved
add hierarchical pretraining from SLIP-YFCCSimCLR · FGSC23 kNNaccuracy: 57.3 -> 55.1Hurt
use location instead of GSD as the only metadata field (ablation)SatMIP · fMoW linear probemacro F1: 37.4 -> 47.1Improved
add timestamp and 12 other fields to GSD and location (ablation)SatMIP · fMoW linear probemacro F1: 51.3 -> 50.7No clear effect
use strong SimCLR augmentation instead of light random resized crop (ablation)SatMIP · fMoW kNNmacro F1: 44.6 -> 45.6Improved
couple one image view between SimCLR and SatMIP objectives (ablation)SatMIPS · fMoW kNNtraining time (rel.): 1.53 -> 1.05Improved
train with SatMIPS instead of SimCLR for faster convergenceSimCLR · all downstream datasetskNN performanceImproved

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