Image colorization has been studied to obtain color images from monochrome images. Recently, deep-learning-based colorization algorithms have enabled the automatic colorization of an image. In this manuscript, we describe some image colorization algorithms, which can be classified into three kinds on the basis of the difference in the intervention by the user: some color scribbles by the user, the source color image provided by the user, and no intervention of the user. Furthermore, colorization applications, such as medical image processing, animation production, and color image coding, are also described.
View full abstract