The Pytorch implementation for: “EfficientCD: A New Strategy For Change Detection Based With Bi-temporal Layers Exchanged[]([2407.15999/] EfficientCD: A New Strategy For Change Detection Based With Bi-temporal Layers Exchanged (arxiv.org)), Sijun Dong, Yuwei Zhu, Geng Chen, Xiaoliang Meng::yum::yum:

[EfficientCD](EfficientCD: A New Strategy For Change Detection Based With Bi-temporal Layers Exchanged | IEEE Journals & Magazine | IEEE Xplore) has been accepted in [IEEE TGRS](IEEE Xplore: IEEE Transactions on Geoscience and Remote Sensing)

Requirement

env.yaml

Revised parameters

check the configs

Training, Test and Visualization Process

bash tools/train.sh

EfficientCD Pretrained Weights And Test Results

LEVIR-CD: 链接：https://pan.baidu.com/s/1epOgO-cw1gDsLdKwnb_Etw 提取码：k7hu

(This experimental setting is different from the experimental setting description of the LEVIR-CD dataset in the original paper. It adopts the same experimental setting method as the CLCD dataset, using random cutting training and sliding window prediction.)

WHUCD: 链接：https://pan.baidu.com/s/12_O_CdDemhidzNw1jJUwCA 提取码：u1md

CLCD: 链接: https://pan.baidu.com/s/1Ha4VR2KNhY0Mi7uaFinmWQ 提取码: viqe

Citation

If you use this code for your research, please cite our papers.

@ARTICLE{10608163,
  author={Dong, Sijun and Zhu, Yuwei and Chen, Geng and Meng, Xiaoliang},
  journal={IEEE Transactions on Geoscience and Remote Sensing}, 
  title={EfficientCD: A New Strategy For Change Detection Based With Bi-temporal Layers Exchanged}, 
  year={2024},
  volume={},
  number={},
  pages={1-1},
  keywords={Feature extraction;Remote sensing;Task analysis;Computational modeling;Transformers;Biological system modeling;Land surface;Change detection;feature interaction;Euclidean distance},
  doi={10.1109/TGRS.2024.3433014}}

Acknowledgments

Our code is inspired and revised by open-mmlab/mmsegmentation, timm. Thanks for their great work!!