Grayscale Image Colorization Using Deep Learning and OpenCV

Grayscale image colorization is the process of predicting and assigning accurate colours to given grayscale image with an aim to recreate a realistic coloured image. The method proposed in this paper uses OpenCV and deep learning with CNN (Convolutional neural network) and GAN (Generative adversarial networks). Here CNN is used for understanding and mapping the relation between grayscale and colour images, GAN is used for generating realistic images by improving the visual quality using generator and discriminator and OpenCV is used for accelerating machine perception and providing a common ground for computer vision applications. The output of this method produces results with better accuracy and visual quality than the existing methods therefore acting as an effective and efficient solution for grayscale image colourization.