Home

Awesome

Deep Metric Learning for Unsupervised CD

Official PyTorch implementation of "Deep Metric Learning for Unsupervised Change Detection in Remote Sensing Images" (Metric-CD)

paper: https://arxiv.org/abs/2303.09536

Introduction

<p float="left"> <img src="/imgs/img1.png" width="200" /> <img src="/imgs/img2.png" width="200" /> <img src="/imgs/cm.png" width="200" /> <img src="/imgs/lasvegas.gif" width="200" /> </p> Figure 1. Demonstration of how our proposed algorithm finds changes in the multi-temporal remote sensing image (for beirut image pair in OSCD dataset).

Demo on OSCD dataset

Conda environment.

Create a virtual conda environment using the provided environment.yml (you may use: conda env create -f environment.yml).

Run Demo.

Once the environment is setup, open Demo_OSCD from jupyter notebook and simply run the file.