MAFF-Net: Multi-Attention Guided Feature Fusion Network for Change Detection in Remote Sensing Images

Abstract

One of the most important tasks in remote sensing image analysis is remote sensing image Change Detection (CD), and CD is the key to helping people obtain more accurate information about changes on the Earth’s surface. A Multi-Attention Guided Feature Fusion Network (MAFF-Net) for CD tasks has been designed. The network enhances feature extraction and feature fusion by building different blocks. First, a Feature Enhancement Module (FEM) is proposed. The FEM introduces Coordinate Attention (CA). The CA block embeds the position information into the channel attention to obtain the accurate position information and channel relationships of the remote sensing images. An updated feature map is obtained by using an element-wise summation of the input of the FEM and the output of the CA. The FEM enhances the feature representation in the network. Then, an attention-based Feature Fusion Module (FFM) is designed. It changes the previous idea of layer-by-layer fusion and chooses cross-layer aggregation. The FFM is to compensate for some semantic information missing as the number of layers increases. FFM plays an important role in the communication of feature maps at different scales. To further refine the feature representation, a Refinement Residual Block (RRB) is proposed. The RRB changes the number of channels of the aggregated features and uses convolutional blocks to further refine the feature representation. Compared with all compared methods, MAFF-Net improves the F1-Score scores by 4.9%, 3.2%, and 1.7% on three publicly available benchmark datasets, the CDD, LEVIR-CD, and WHU-CD datasets, respectively. The experimental results show that MAFF-Net achieves state-of-the-art (SOTA) CD performance on these three challenging datasets.

DAHT-Net: Deformable Attention-Guided Hierarchical Transformer Network Based on Remote Sensing Image Change Detection

“…By adding a transformer encoder to the CNN backbone network, BIT-CD models the context in a compact token-based spacetime. 5) STNet [32] adopts a fusion of self-encoder and 3D CNN to enhance change detection while maintaining a lightweight and deployable model for various fields. However, extensive data is necessary for effective training, and the model structure is relatively intricate, requiring specialized technical expertise for design and implementation.…”

Building Change Detection by W-Shape Resunet++ Network With Triple Attention Mechanism

“…The use of attention mechanisms is one of the ways of extracting distinguishing features in deep network. Channel attention (Liu et al, 2020), spatial attention or both of them (Ding et al, 2021;Ma et al, 2022) are used to improve the performance of CNNs. These attention mechanisms have improved representation of the features generated by standard convolutional layers.…”

CCSNet: a cascaded cross-scale network for remote sensing image change detection

“…Cheng et al 42 inserted channel attention into the Siamese encoder backbone to extract specific semantic features and used spatial attention to enhance the positional change response. Ma et al 43 introduced coordinate attention to obtaining more accurate position information and channel relationship. Pan et al 44 used convolutional block attention modules (CBAM) to learn channel and spatial information in features, improving the model's ability to locate the identical locations between bitemporal images and determine changes in small objects.…”

“…inserted channel attention into the Siamese encoder backbone to extract specific semantic features and used spatial attention to enhance the positional change response. Ma et al 43 . introduced coordinate attention to obtaining more accurate position information and channel relationship.…”