dual unet:a novel siamese network for change detection with cascade differential fusion

TL;DR

This paper introduces Dual-UNet, a Siamese neural network with differential-attention and multi-scale fusion for improved change detection in remote sensing images, outperforming existing methods.

Contribution

It presents a novel Siamese network architecture with differential-attention modules and a multi-scale fusion strategy for enhanced change detection accuracy.

Findings

Outperforms state-of-the-art methods on seasonal change detection datasets

Uses differential-attention to focus on pixel-wise spatial differences

Employs multi-scale weighted variance map fusion for better feature integration

Abstract

Change detection (CD) of remote sensing images is to detect the change region by analyzing the difference between two bitemporal images. It is extensively used in land resource planning, natural hazards monitoring and other fields. In our study, we propose a novel Siamese neural network for change detection task, namely Dual-UNet. In contrast to previous individually encoded the bitemporal images, we design an encoder differential-attention module to focus on the spatial difference relationships of pixels. In order to improve the generalization of networks, it computes the attention weights between any pixels between bitemporal images and uses them to engender more discriminating features. In order to improve the feature fusion and avoid gradient vanishing, multi-scale weighted variance map fusion strategy is proposed in the decoding stage. Experiments demonstrate that the proposed approach consistently outperforms the most advanced methods on popular seasonal change detection datasets.