AI Paper Atlas

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

TerraFM: A Scalable Foundation Model for Unified Multisensor Earth Observation

Muhammad Sohail Danish, Muhammad Akhtar Munir, Syed Roshaan Ali Shah, Muhammad Haris Khan, Rao Muhammad Anwer, Jorma Laaksonen, Fahad Shahbaz Khan, Salman Khan

ICLR 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 18 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 modality as augmentation (ablation)TerraFM-B · m-EuroSat/m-BigEarthNetkNN accuracy: 83.20 -> 87.70 (EuroSat); 54.62 -> 57.63 (BigEarthNet)Improved
add cross-attention fusion (ablation)TerraFM-B · m-Cashew-PlantmIoU: 59.17 -> 62.40Improved
add cross-attention fusion (ablation)TerraFM-B · m-BigEarthNetkNN F1: 57.63 -> 57.74No clear effect
add dual centering (ablation)TerraFM-B · m-EuroSat/m-Cashew-PlantkNN accuracy / mIoU: 88.50 -> 90.40 (EuroSat); 62.40 -> 64.58 (Cashew)Improved
use global cross-attention with 196 queries instead of Q=5 spatial fusion (ablation)TerraFM-B · m-EuroSatkNN accuracy: 90.4 -> 77.1Hurt
replace convolutional patch embedding with ViT-S token extraction (ablation)TerraFM-B · m-EuroSatkNN accuracy: 90.4 -> 87.2Hurt
use more spatial queries (Q=5 vs Q=1) (ablation)TerraFM-B · m-EuroSatkNN accuracy: 89.2 -> 90.4Improved
use TerraFM instead of Galileo/CROMA baselinesTerraFM · Copernicus-Bench DFC2020-S2mIoU: CROMA 66.5 vs TerraFM 63.8Hurt
scale TerraFM-B to TerraFM-LTerraFM · m-Brick-Kiln kNNaccuracy: 94.5 -> 93.0Hurt
use TerraFM-120M instead of Prithvi-EO-2.0 300MTerraFM · Landslide4SensemIoU: 65.0 -> 70.8Improved

Later papers that compared with this method

2 later accepted papers report TerraFM as a baseline in their results tables; 2 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 0 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