Artificial Intelligent approach for Colorful Image Colorization Using a DCNN

Abstract

Coloring grayscale photos manually or using traditional coloring methods takes extensive user interaction. This may involve applying many colored scribbles, viewing related images, or doing segmentation. Even the most sophisticated software available in this day and age can take up to a month to color an image that was originally black and white. This occurs because the image contains a wide variety of color tones and tints. In this research work, we offer an innovative method for colorizing grayscale photographs that use deep learning techniques. First, we can separate the subject matter and aesthetic of several images and then recombine them into a single image by using a pre-trained convolutional neural network that was first developed for image categorization. Following this, we present an approach that may colorize a black-and-white image by combining the content of the black-and-white image with the style of a color image that has semantic similarities with the black-and-white image.