Image Colorization with Generative Adversarial Networks

Over the last decade, the process of automatic colorization had been studied thoroughly due to its vast application such as colorization of grayscale images and restoration of aged and/or degraded images. This problem is highly ill-posed due to the extremely large degrees of freedom during the assignment of color information. Many of the recent developments in automatic colorization involved images that contained a common theme throughout training, and/or required highly processed data such as semantic maps as input data. In our approach, we attempted to fully generalize this procedure using a conditional Deep Convolutional Generative Adversarial Network (DCGAN). The network is trained over datasets that are publicly available such as CIFAR-10 and Places365. The results between the generative model and tradition deep neural networks are compared.