Gray Scale Image Colorization using Convolutional Neural Network and PyTorch

Colorizing grayscale photos is a difficult process that has important uses in the creative industries, media improvement, and historical photo restoration. By utilizing advances in neural networks for image-to-image translation, this project investigates colorization of black-and-white images using PyTorch and deep learning approaches. The technique uses a convolutional neural network (CNN) architecture trained on a range of colourful image datasets to predict chrominance components while maintaining brightness from the input grayscale image. Notably, that the model outperforms in the conventional techniques in terms of evaluation measures like PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural Similarity Index), achieving the good realistic and brilliant colorization. In addition to advancing automated image improvement, this research offers useful applications for repairing old photos, improving visual media, and supporting the creative industries by colorizing the black and white images.