Towards a satellite image manipulation and deepfake localization benchmark dataset

2026-08-05

AI Summary

This study addresses the critical lack of high-quality, fine-grained satellite image manipulation datasets in remote sensing, which has hindered the development of forgery detection and localization algorithms. To bridge this gap, the authors introduce the first benchmark dataset specifically designed for satellite image tampering detection, comprising 30 manipulated images—altered via copy-paste splicing and diffusion-based inpainting—and 30 authentic images. Each image is accompanied by pixel-level tampering masks and comprehensive acquisition metadata. This dataset uniquely enables pixel-level localization evaluation and facilitates systematic investigation into how image acquisition parameters influence detection performance. The dataset has been publicly released on the Hugging Face platform, providing foundational support for research in geospatial deepfakes and image forensics.

Abstract

Verifying the authenticity of satellite imagery has become increasingly critical given advances in generative artificial intelligence. Highly realistic synthetic imagery produced for malicious purposes (deepfakes) can have major consequences in the remote sensing domain, where this data is a fundamental source of information for science applications, planning, logistics, and monitoring. The remote sensing community lacks high-quality, fine-grained manipulation datasets suitable for training and evaluating detection and image forensics algorithms. Existing datasets are lacking and those that do exist either provide no ground truth masks for evaluating manipulation localization, or consist of entire images generated by GANs or diffusion models, which are inadequate for measuring localization performance. To address this gap, we describe a preliminary dataset construction process and prototype benchmark dataset for satellite image manipulation detection and localization. The dataset contains 60 images total, with 30 images carefully manipulated using three manipulation types including copy-paste splicing and diffusion model inpainting, and 30 authentic images. Each image is accompanied by a ground-truth mask and acquisition metadata, enabling both pixel-level localization metrics, image metadata studies, and analyses of how manipulation detection performance relates to image collection parameters. We describe the dataset construction process and present this initial release to support further research in image forensics and geospatial deepfake detection. The prototype dataset can be downloaded at https://huggingface.co/datasets/geodf/fmow-fake-small.

Problem

• satellite image manipulation
• deepfake localization
• image forensics
• benchmark dataset
• remote sensing

Innovation

• satellite image forensics
• deepfake localization
• manipulation dataset
• ground-truth mask
• diffusion inpainting