Efficient Geometry-Controlled High-Resolution Satellite Image Synthesis

TL;DR

This paper introduces a simple method to control high-resolution satellite image synthesis using pre-trained diffusion models, improving geometric alignment with minimal additional complexity.

Contribution

It proposes a novel control technique leveraging skip connection features and windowed cross-attention, enhancing geometry control in satellite image synthesis.

Findings

Method achieves comparable performance to existing controls

Better alignment with geometry control map

Highlights limitations in current evaluation methods

Abstract

High-resolution satellite images are often scarce and costly, especially for remote areas or infrequent events. This shortage hampers the development and testing of machine learning models for land-cover classification, change detection, and disaster monitoring. In this paper, we tackle the problem of geometry-controlled high-resolution satellite image synthesis by adding control over existing pre-trained diffusion models. We propose a simple yet efficient method for controlling the synthesis process by leveraging only skip connection features using windowed cross-attention modules. Several previously established control techniques are compared, indicating that our method achieves comparable performance while leading to a better alignment with the geometry control map. We also discuss the limitations in current evaluation approaches, amplifying the necessity of a consistent alignment assessment.