This paper presents an enhanced deep super-resolution network (EDSR) and a multi-scale deep super-resolution system (MDSR) for single image super-resolution (SISR). The authors optimize the conventional residual network by removing unnecessary modules, improving computational efficiency and performance. They also introduce a multi-scale architecture that shares parameters across different scales, reducing model size while maintaining comparable performance. The proposed methods outperform state-of-the-art techniques on benchmark datasets and win the NTIRE2017 Super-Resolution Challenge. The key contributions include a simplified residual network structure, the use of residual scaling for stable training, and a multi-scale model that leverages inter-scale correlations. The models achieve superior performance in terms of PSNR and SSIM, demonstrating their effectiveness in reconstructing high-resolution images from low-resolution inputs.This paper presents an enhanced deep super-resolution network (EDSR) and a multi-scale deep super-resolution system (MDSR) for single image super-resolution (SISR). The authors optimize the conventional residual network by removing unnecessary modules, improving computational efficiency and performance. They also introduce a multi-scale architecture that shares parameters across different scales, reducing model size while maintaining comparable performance. The proposed methods outperform state-of-the-art techniques on benchmark datasets and win the NTIRE2017 Super-Resolution Challenge. The key contributions include a simplified residual network structure, the use of residual scaling for stable training, and a multi-scale model that leverages inter-scale correlations. The models achieve superior performance in terms of PSNR and SSIM, demonstrating their effectiveness in reconstructing high-resolution images from low-resolution inputs.