2025.1 our paper has been accepted by IEEE TIP.

2024.2 arxiv paper released.

you can find them here.

you can download here

During development, we referred to the following resources:

• MAE - The backbone of SENet is built based on this project. Note that to get good results, you need to use the full pre-trained weights of MAE (including the decoder part). you can get the mae weight here.
• BGNet - The training process and implementation are based on this project.