CEBSNet: Change-Excited and Background-Suppressed Network with Temporal Dependency Modeling for Bitemporal Change Detection

2025-05-21

AI Summary

Addressing challenges in bitemporal remote sensing and street-view change detection—including illumination/seasonal variations, background clutter, and missed detection of subtle changes over large temporal intervals—this paper proposes a novel framework that jointly models temporal dependencies and multi-scale perception. Our method introduces three key innovations: (1) a Channel Swapping Module (CSM) to explicitly capture cross-temporal feature dependencies; (2) a synergistic Feature Excitation and Background Suppression Mechanism (FESM) to enhance sensitivity to subtle changes while suppressing irrelevant background; and (3) a Pyramid-Aware Spatial–Semantic Attention module (PASCA) to preserve structural integrity of change regions while focusing on salient areas. Evaluated on three street-view and two remote sensing benchmarks, our approach consistently outperforms state-of-the-art methods, achieving significant gains in both subtle-change detection recall and pixel-level localization accuracy.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Change detection, a critical task in remote sensing and computer vision, aims to identify pixel-level differences between image pairs captured at the same geographic area but different times. It faces numerous challenges such as illumination variation, seasonal changes, background interference, and shooting angles, especially with a large time gap between images. While current methods have advanced, they often overlook temporal dependencies and overemphasize prominent changes while ignoring subtle but equally important changes. To address these limitations, we introduce extbf{CEBSNet}, a novel change-excited and background-suppressed network with temporal dependency modeling for change detection. During the feature extraction, we utilize a simple Channel Swap Module (CSM) to model temporal dependency, reducing differences and noise. The Feature Excitation and Suppression Module (FESM) is developed to capture both obvious and subtle changes, maintaining the integrity of change regions. Additionally, we design a Pyramid-Aware Spatial-Channel Attention module (PASCA) to enhance the ability to detect change regions at different sizes and focus on critical regions. We conduct extensive experiments on three common street view datasets and two remote sensing datasets, and our method achieves the state-of-the-art performance.

Problem

• Detect pixel-level changes in bitemporal images despite illumination and seasonal variations
• Model temporal dependencies to reduce noise and improve change detection accuracy
• Enhance detection of both prominent and subtle changes in varied image conditions

Innovation

• Channel Swap Module models temporal dependency
• Feature Excitation and Suppression Module captures changes
• Pyramid-Aware Spatial-Channel Attention enhances detection