Image colorization is a fascinating and challenging topic in image-to-image translation. Image Colorization is the process of coloring a grayscale image by using a black and white image known as a grayscale image as input and obtaining the output in RGB format simply known as a color image. The fundamental aim is to convince the spectator that the outcome is genuine. Many cameras, such as surveillance cameras and satellite cameras, still capture grayscale images which are kind of hard to analyze. Over the last 20 years, a wide range of colorization methods have been created, ranging from algorithmically simple yet time- and energy-consuming procedures due to unavoidable human participation to more difficult but also more automated methods. Image colorization by hand is time-consuming and prone to human error. Automatic conversion has evolved into a difficult field that combines machine learning and deep learning with art. In this project, we built an image colorization framework based on a deep learning method known as Generative Adversarial Networks, or GANs for short. GANs are a type of generative modelling that employs deep learning techniques such as convolutional neural networks