The code and data here are for the following paper:
Kewei Liu, Zhenbo Ren, Xiaoyan Wu, Jianglei Di, Jianlin Zhao, "SSG-Net: A Robust Network for Adaptive Multi-Source Image Registration Based on SuperGlue," Digital Signal Processing, under review.
Feel free to let me know if you have any problems at zbrenATnwpu.edu.cn.
The code provided here is to train a network for multimodal image registration. For a generalized utilization to the public, the code is given in Python and PyTorch and for GPU use. For demonstration purposes, adequate minor revisions are made in the code.
Associated data can be found at https://drive.google.com/file/d/1wx_zOzmQj7HkqVMJNkwkoOenwDMueHpj/view?usp=sharing.
Weights can be found at https://drive.google.com/file/d/1LG7-5l-KeDJaYNFdBL9wH_sig9JfTrvi/view?usp=sharing, and they should be put in the folder "models".
Modules that the code depends on are given in the first functional block.