Image restoration, a primary task or work in computer vision and image processing, which seeks to restore damaged or deteriorated photographs to their original, high-quality state. This paper presents an innovative approach to image restoration utilizing Generative Adversarial Networks (GANs). GANs have become famous because they are able to generate realistic data samples, which make them well suited for the task of image restoration, a very challenging task. In recent and present years Generative Adversarial Networks (GANs) have become prominent as a powerful approach for image restoration tasks, demonstrating significant advancements in generating high-quality and realistic images. GANs are a type of deep learning models which consists of two neural networks, the generator and the discriminator, they compete with each other to inspect closely and thoroughly, capture, and replicate the variations within a data- set. The generator learns to synthesize images, as the discriminator gains the ability to discriminate between produced and genuine images. Through the process of adversarial training, the generator progressively improves its ability to generate realistic images, culminating in amplified image restoration skills.