AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

MMEarth: Exploring Multi-Modal Pretext Tasks For Geospatial Representation Learning

Vishal Nedungadi, Ankit Kariryaa, Stefan Oehmcke, Serge Belongie, Christian Igel, Nico Lang

ECCV 2024 · Vision-language and multimodal › Cross-modal learning · approach: Multi-modality image fusion · 69 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
replace ImageNet pretraining data with Sentinel-2 RGB at patch size 32 (ablation)ConvNeXt V2 MAE · BigEarth20k / SAcrop3k fine-tuningF1 / IoU: SAcrop3k 26.7 -> 23.8; BigEarth20k FT 55.7 -> 55.5No clear effect
reduce patch size from 32 to 16 for 112px satellite images (ablation)MAE on Sentinel-2 RGB · So2Sat20k fine-tuningaccuracy: 38.1 -> 48.8Improved
use all 12 Sentinel-2 bands instead of RGB only (ablation)MAE on Sentinel-2 · So2Sat20k fine-tuningaccuracy: 54.1 -> 48.7Hurt
use all 12 Sentinel-2 bands instead of RGB only (ablation)MAE on Sentinel-2 · SAcrop3k fine-tuningIoU: 33.1 -> 36.0Improved
add multi-modal pretext tasks to Sentinel-2 reconstruction (ablation)MP-MAE · BigEarth20k linear probingF1: 36.2 -> 40.9Improved
add multi-modal pretext tasks to Sentinel-2 reconstruction (ablation)MP-MAE · So2Sat20k linear probingaccuracy: 33.5 -> 41.7Improved
add only pixel-level or only image-level pretext tasks (ablation)MP-MAE · BigEarth20k linear probingF1: 36.2 -> 36.6No clear effect
combine pixel-level and image-level pretext groups (ablation)MP-MAE · BigEarth20k linear probingF1: 36.6 -> 40.9Improved
multi-modal pretext tasks in few-shot linear probingMP-MAE · EuroSat2k 1% of labelsaccuracy: 16.5% -> 34%Improved
MP-MAE (3.7M encoder) versus ResNet-50 (25M) baselines on linear probing BigEarth20kMP-MAE · BigEarth20k linear probingF1: 44.5 -> 43.3 (vs SeCo)Hurt

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