Image Colorization
Image colorization aims to automatically add realistic color to grayscale images or videos, a challenging problem due to inherent ambiguity. Current research focuses on improving color accuracy and consistency, often employing generative adversarial networks (GANs), transformers, and diffusion models, sometimes incorporating additional modalities like text descriptions or audio to guide the process. These advancements have implications for various fields, including digital image restoration, historical preservation, and enhancing the capabilities of other imaging technologies like LiDAR. The ongoing emphasis is on achieving greater controllability, realism, and efficiency in colorization algorithms.
Papers
Analysis of Different Losses for Deep Learning Image Colorization
Coloma Ballester, Aurélie Bugeau, Hernan Carrillo, Michaël Clément, Rémi Giraud, Lara Raad, Patricia Vitoria
Influence of Color Spaces for Deep Learning Image Colorization
Coloma Ballester, Aurélie Bugeau, Hernan Carrillo, Michaël Clément, Rémi Giraud, Lara Raad, Patricia Vitoria