Skip to content

jiaerfei/YOLO-SimSiam

Repository files navigation

YOLO-SimSiam

A self-supervised learning-based method for tunnel leakage defect few-shot recognition and visual explanation

workflow

1. 环境配置

参考ultralytics, Grad-CAM等 (https://github.com/ultralytics/ultralytics) (https://github.com/jacobgil/pytorch-grad-cam) 安装所需的各类第三方库

2. 文件说明

SSL_main.py

用于实现图像的特征学习

visualize.py

用于可视化各类结果,包含用于对比学习的图像,对比特征图,分割特征图等

box_main.py

用于正常训练和少样本训练

utils.py

用于数据预处理,标签格式转换等

其他为模型、数据集、模型最优权重等配置文件

3. 数据集下载

Leakage.zip 链接: https://pan.baidu.com/s/1SeBuYLQCoXsSWbdvB8urmA?pwd=2025

About

A self-supervised learning-based method for tunnel leakage defect few-shot recognition and visual explanation

Resources

License

Stars

1 star

Watchers

1 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages