A Few-shot learning approach for Monkeypox recognition from a cross-domain perspective

This repository contains the reference source code for the paper A Few-shot learning approach for Monkeypox recognition from a cross-domain perspective published in Journal of Biomedical Informatics.

Directory Structure

./dataset contains the code used to generate datasets.
./files contains the used image data.
./fine_tuning contains metric-based few-shot learning methods.
./metrics contains transfer-based few-shot learning methods.

Prerequisites

• Python 3.10
• Pytorch, Torchvision, Numpy

Getting started

Installation

• Clone this repository: git clone https://github.com/000linlin/SCA.git
• Install PyTorch and other dependencies.

Train

• Run python ./fine_tuning/train.py to train the SCA and other transfer-based few-shot learning methods.

• Run python ./metrics/train.py to train the metric-based few-shot learning methods.

Test

• Run python ./fine_tuning/test.py to test the task of recognizing chickenpox, measles, and human monkeypox diseases by SCA and others in the few-shot standard.

• Run python /metrics/test.py to test the metric-based few-shot learning methods in this task.

Citation

If you use this code, please consider citing our paper:

@article{chen2023few,
  title={A Few-shot learning approach for Monkeypox recognition from a cross-domain perspective},
  author={Chen, Bolin and Han, Yu and Yan, Lin},
  journal={Journal of Biomedical Informatics},
  pages={104449},
  year={2023},
  publisher={Elsevier}
}