Work

A Transformer-Based Siamese Network for Change Detection

Year: 2022

Type: conference-paper

Abstract: This paper presents a transformer-based Siamese network architecture (abbreviated by ChangeFormer) for Change Detection (CD) from a pair of co-registered remote sensing images. Different from recent CD frameworks, which are based on fully convolutional networks (ConvNets), the proposed method unifies hierarchically structured transformer encoder with Multi-Layer Perception (MLP) decoder in a Siamese network architecture to efficiently render multi-scale long-range details required for accurate CD. Experiments on two CD datasets show that the proposed end-to-end trainable ChangeFormer architecture achieves better CD performance than previous counterparts. Our code and pre-trained models are available at github.com/wgcban/ChangeFormer.

Source: IEEE International Geoscience and Remote Sensing Symposium (IGARSS)

DOI: 10.1109/igarss46834.2022.9883686

Authors: Wele Gedara Chaminda Bandara, Vishal M. Patel

Institution: Johns Hopkins University

Language: English

Topic: Remote-Sensing Image Classification

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funder: National Science Foundation

Award: CAREER: Seeing Through Atmospheric Turbulence: Image Restoration and Understanding using Deep Convolutional Neural Networks