CS-Diffformer: A Change-Sensitive Transformer for Remote Sensing Change Detection

Change detection (CD) in remote sensing imagery plays a pivotal role in applications such as urban development monitoring, disaster assessment, and environmental surveillance. While deep learning methods based on convolutional neural networks (CNNs) and Transformers have advanced the field, existing models still face limitations in feature alignment, sensitivity to fine-grained changes, and reliance on pre-trained weights. To address these challenges, we propose CS-DiffFormer (Change-Sensitive Diffusion-based Transformer), a novel Transformer-based framework tailored for change detection tasks. CS-DiffFormer integrates two key modules: the Dynamic Sparse Self-Attention (DSSA) and the Dynamic Differential Feature Fusion (DDFF). These modules jointly enhance the representation of semantic changes and suppress irrelevant noise. Experimental results on four public datasets demonstrate that CS-DiffFormer achieves state-of-the-art performance, validating its effectiveness and generalization capability.