AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Can Generative Geospatial Diffusion Models Excel as Discriminative Geospatial Foundation Models?

Yuru Jia, Valerio Marsocci, Ziyang Gong, Xue Yang, Maarten Vergauwen, Andrea Nascetti

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Diffusion models as perception tools · 9 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 learnable fusion (global weighted) instead of simple concatenation (ablation)SatDiFuser · m-cashew-plantationmIoU: 59.1 -> 66.5Improved
add learnable fusion (global weighted) instead of simple concatenation (ablation)SatDiFuser · m-eurosattop-1 accuracy: 94.5 -> 97.7Improved
add learnable fusion (global weighted) instead of simple concatenation (ablation)SatDiFuser · m-so2sattop-1 accuracy: 55.4 -> 59.3Improved
use multi-timestep multi-block features with fusion instead of a single timestep-1 self-attention block (ablation)SatDiFuser · m-so2sattop-1 accuracy: 53.6 -> 59.3Improved
use two experts activated (top-k 2) instead of one (ablation)SatDiFuser MoE fusion · m-chesapeake-landcovermIoU: 60.2 -> 71.6Improved
increase experts from 8 to 12 (ablation)SatDiFuser MoE fusion · m-chesapeake-landcovermIoU: 71.6 -> 68.3Hurt
reduce experts from 8 to 4 (ablation)SatDiFuser MoE fusion · m-chesapeake-landcovermIoU: 71.6 -> 69.8Hurt
use satellite-finetuned DiffusionSat instead of Stable Diffusion v2-1 backbone (ablation)SatDiFuser Global fusion · m-chesapeake-landcovermIoU: 65.8 -> 69.5Improved
use DiffusionSat instead of SD v2-1 backbone (ablation)SatDiFuser Global fusion · m-cashew-plantationmIoU: 63.7 -> 66.5Improved
add SatDiFuser global fusion versus best baseline with 10% labelsSatDiFuser vs CROMA · m-cashew-plantation (10% labels)mIoU: 34.3 -> 39.6Improved
add SatDiFuser MoE fusion versus best pretrained modelSatDiFuser vs CROMA · m-bigearthnetF1: 58.3 -> 65.8Improved
use sampling within first 20% of diffusion timesteps (ablation)SatDiFuser features · m-eurosat and m-NeonTreeaccuracy / mIoUImproved

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