Automatic Colorization and Restoring of gray scale images using Deep Learning

Color plays an important role in our day-to-day life. If the colors are used proper, it can results in improved health zone, mind relief also for humans. When used in the right ways, color can save on energy consumption also. The main goal of this work is to bring new life to old photo and videos by colorizing them. The proposed deep neural network has a fusion layer that allows us to elegantly merge local information dependent on small image patches with global priors computed using the entire image, Furthermore, proposed architecture can process images of any resolution, unlike most existing approaches based on CNN. An existing large-scale scene database is used to train the proposed model to learn the global priors and classify the objects of image to be able to map color. It demonstrates method extensively on many different types of images, including black-and-white photography from over a hundred years ago, and show realistic colorizations. The colorization of our proposed approach is considered “natural” more than 80% for images.