HSANET: A Hybrid Self-Cross Attention Network For Remote Sensing Change Detection

2025-04-21

AI Summary

Remote sensing image change detection faces challenges including insufficient multi-scale feature extraction, weak global and cross-scale contextual modeling, and low accuracy in edge-detail identification. To address these, we propose HSANet: (1) a hierarchical convolutional architecture constructs a multi-scale feature pyramid; (2) the first Hybrid Self-Cross Attention (HSCA) module enables dynamic cross-scale feature alignment and edge-response enhancement without auxiliary supervision; and (3) a multi-scale feature fusion mechanism jointly optimizes detection of small objects and boundary changes. Evaluated on multiple public benchmarks—including LEVIR-CD and WHU-CD—HSANet achieves state-of-the-art performance, significantly improving overall accuracy and edge completeness. The source code is publicly available.

Technology Category

• Computer Vision: Object Detection & Categorization
• Machine Learning: Learning on the Edge & Model Compression
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Cloud, edge and content delivery systems for the Web
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

The remote sensing image change detection task is an essential method for large-scale monitoring. We propose HSANet, a network that uses hierarchical convolution to extract multi-scale features. It incorporates hybrid self-attention and cross-attention mechanisms to learn and fuse global and cross-scale information. This enables HSANet to capture global context at different scales and integrate cross-scale features, refining edge details and improving detection performance. We will also open-source our model code: https://github.com/ChengxiHAN/HSANet.

Problem

• Detect changes in remote sensing images accurately
• Extract multi-scale features using hierarchical convolution
• Fuse global and cross-scale information via attention mechanisms

Innovation

• Uses hierarchical convolution for multi-scale features
• Incorporates hybrid self and cross-attention mechanisms
• Captures global context and refines edge details