utilizes state-of-the-art deep learning models and adjusts parameters according to quality metrics such as structural similarity (SSIM) and perceptual loss to create an image that is indistinguishable from a real picture