Automatic Image Colorization using Generative Adversarial Networks

Image colorization is an approach of transforming a black and white image into colorized image. The colonization process can also be used to perform color corrections. This application has been incorporated in large software like Adobe Photoshop, After Effects and Lightroom, and Da Vinci Resolve to aid users through their editing process. In the past, the process of colorization required a tremendous amount of human involvement and the results were still not properly saturated. The approach considered for this topic is a fully generalized procedure using a conditional Deep Convolutional Generative Adversarial Networks (DCGAN). Using datasets like Cifar-10 and MNIST , the model was trained.