For this purpose, in this article, we use the stable encoder- decoder architecture, combined with a grid-based attention gate and atrous spatial pyramid pooling module, to capture and restore features ...