HRFNet: High-Resolution Forgery Network for Localizing Satellite Image Manipulation

TL;DR

HRFNet is a novel high-resolution network designed for precise satellite image forgery localization, effectively integrating multi-scale features to improve accuracy without increasing computational costs.

Contribution

The paper introduces HRFNet, a high-resolution model that combines shallow and deep features for improved forgery localization in satellite images, overcoming limitations of previous patch-based methods.

Findings

Achieves superior localization accuracy compared to existing methods.

Maintains low memory usage and fast processing speeds.

Effectively integrates RGB and resampling features at multiple scales.

Abstract

Existing high-resolution satellite image forgery localization methods rely on patch-based or downsampling-based training. Both of these training methods have major drawbacks, such as inaccurate boundaries between pristine and forged regions, the generation of unwanted artifacts, etc. To tackle the aforementioned challenges, inspired by the high-resolution image segmentation literature, we propose a novel model called HRFNet to enable satellite image forgery localization effectively. Specifically, equipped with shallow and deep branches, our model can successfully integrate RGB and resampling features in both global and local manners to localize forgery more accurately. We perform various experiments to demonstrate that our method achieves the best performance, while the memory requirement and processing speed are not compromised compared to existing methods.