Researchers have proposed various methods for image restoration, including a multi-stage hybrid attention and adaptive feature fusion network. Other approaches involve multi-attention fusion transformers for single-image super-resolution and image deblurring methods using improved U-Net models with multilayer fusion and attention mechanisms. Additionally, techniques for fine-grained multi-focus image fusion based on edge features have also been explored.