Coloring an Image Using Generative Adversarial Networks (GAN)

Past few years ago there were only gray scale images videos, as technologies has improved those gray scales images are now colorized, which depicts the exact color as it is visible to human eye. Now these colorization techniques are done using deep learning and are automated because of which they have impressive performance. In 2002 a calculation was suggested that colorized pictures through surface combination. Colorization was finished by matching luminance and surface data between a current shading picture and the grayscale picture to be colorized. A technique was proposed as an elective detailing to the colorization issue in 2004. This plan followed a converse methodology, where the expense work was planned by the contrast between each pixel and a weighted normal of its adjoining pixels. To overcome abnormal colors and to improve image quality with good colorization we are using Generative Adversarial Network (GAN). Studies shows that GAN outperforms existing methods