Reviving Black and White Images: Enhancing Colorization with Generative Adversarial Networks (GANs)

One of the more intriguing deep learning applications is colourizing black and white photographs. The restoration of old or damaged images is one application that has drawn a lot of interest in the automatic image colorization technology in the decade since its introduction. Prior to AI and deep learning, this work required a lot of human intervention and hardcoding, but today, the entire process has been entirely automated from beginning to end. Using GANs to restore and recolorize historical photos is one answer to this problem. The objective of this project is to showcase the effectiveness and advantages of GANs Networks by implementing a GAN model capable of transforming fixed-size black and white images into colorized images of identical dimensions. Concerning the dataset, I have approximately 3000 rgb photographs from diverse domains such as mountains, forests, and cities, which we will convert to grayscale and use as labels for our model. I used binary cross entropy as the discriminator's loss function and mean squared error as the generator's loss function, & then I used Adam to optimise the generator and discriminator. Moving on to the results, the colourized output from the generator was significantly closer to the original rgb image.