Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

From Pixels to Patches: Pooling Strategies for Earth Embeddings

Isaac Corley, Caleb Robinson, Inbal Becker-Reshef, Juan M. Lavista Ferres

TL;DR

Generalized Mean Pooling (GeM) is recommended as a drop-in replacement for mean pooling: it improves accuracy without increasing embedding dimensionality.

Abstract

As geospatial foundation models shift from patch-level to pixel-level embeddings, practitioners must aggregate thousands of pixel vectors into patch representations that preserve class-discriminative signal while matching downstream label resolution. The default choice, mean pooling, discards within-patch variability and can drop accuracy by more than 10% under spatial shift. To evaluate this effect, we introduce EuroSAT-Embed: 81,000 embedding GeoTIFFs derived from three foundation models: AlphaEarth, OlmoEarth, and Tessera. We benchmark 11 training-free and 2 parametric pooling methods under both random and geographically disjoint test splits. Our results show that richer pooling schemes reduce the geographic generalization gap by up to 40% relative to mean pooling and increases accuracy by up to 5% on spatial splits. We recommend Generalized Mean Pooling (GeM) as a drop-in replacement for mean pooling: it improves accuracy without increasing embedding dimensionality. For maximum accuracy, Stats pooling (concatenation of min/max/mean/std pooling) performs best at 4x the embedding size. We further find that pooling effectiveness varies across embedding sources and that higher-dimensional embeddings benefit most from distributional statistics.

From Pixels to Patches: Pooling Strategies for Earth Embeddings

TL;DR

Generalized Mean Pooling (GeM) is recommended as a drop-in replacement for mean pooling: it improves accuracy without increasing embedding dimensionality.

Abstract

As geospatial foundation models shift from patch-level to pixel-level embeddings, practitioners must aggregate thousands of pixel vectors into patch representations that preserve class-discriminative signal while matching downstream label resolution. The default choice, mean pooling, discards within-patch variability and can drop accuracy by more than 10% under spatial shift. To evaluate this effect, we introduce EuroSAT-Embed: 81,000 embedding GeoTIFFs derived from three foundation models: AlphaEarth, OlmoEarth, and Tessera. We benchmark 11 training-free and 2 parametric pooling methods under both random and geographically disjoint test splits. Our results show that richer pooling schemes reduce the geographic generalization gap by up to 40% relative to mean pooling and increases accuracy by up to 5% on spatial splits. We recommend Generalized Mean Pooling (GeM) as a drop-in replacement for mean pooling: it improves accuracy without increasing embedding dimensionality. For maximum accuracy, Stats pooling (concatenation of min/max/mean/std pooling) performs best at 4x the embedding size. We further find that pooling effectiveness varies across embedding sources and that higher-dimensional embeddings benefit most from distributional statistics.
Paper Structure (6 sections, 2 figures, 2 tables)

This paper contains 6 sections, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Data and Embeddings
  3. Pooling Methods
  4. Experimental Setup
  5. Results
  6. Discussion

Figures (2)

  • Figure 1: Pixel-to-patch pooling. The input-label resolution mismatch requires aggregating dense pixel embeddings (shown as PCA pseudo-RGB) to a lower resolution for downstream tasks.
  • Figure 2: Random vs spatial accuracy. Points near the diagonal (where random $=$ spatial) have smaller generalization gaps.