MSFIN

This package contains the source code which is associated with the following paper:

Yu Liu, Lei Wang, Juan Cheng, Xun Chen, "Multi-scale Feature Interactive Network for Multi-focus Image Fusion", submitted to IEEE Transactions on Instrumentation and Measurement, 2021. (under review, second round)

Edited by Yu Liu.

Usage of this code is free for research purposes only.

The MSFIN model is learned by pytorch. The training code will be released if the paper could be accepted for publication.

Thank you.

Requirements:

CUDA 10
Python 3.6
Torch 1.8.1
Torchvison 0.9.1
numpy 1.19.4
pydensecrf 1.0

Test :

Put test image pairs in the "sourceimages" folders, and run "test.py" to test the trained model. You can also directly use the trained model we provide.

Contact:

Don't hesitate to contact me if you meet any problems when using this code.

Yu Liu Department of Biomedical Engineering Hefei University of Technology
Email: yuliu@hfut.edu.cn; lyuxxz@163.com Homepage: https://sites.google.com/site/yuliu316316; https://github.com/yuliu316316

Acknowledgement:

The author would like to thank Mr Shuang Xu from Xi’an Jiao Tong University [1] and Mr Xingchen Zhang from Imperial College London [2] for providing the MFFW dataset.

References:

[1] S. Xu, X. Wei, C. Zhang, J. Liu and J. Zhang. MFFW: A new dataset for multi-focus image fusion. arXiv, 2020.

[2] X. Zhang, “Deep learning-based multi-focus image fusion: A survey and a comparative study,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021.