A review of image colourisation

This chapter reviews the recent development of image colourisation, which aims at adding colour to a given greyscale image. There are numerous applications involving image colourisation, such as converting black and white photos or movies to colour, restoring historic photographs to improve the aesthetics of the image, as well as colourising many other types of images lacking colour (eg medical images, infrared night time images). According to the source where the colours come from, the existing methods can be categorised into three classes: colourisation by reference, colourisation by scribbles and colourisation by deep learning. In this chapter, we introduce the basic idea and survey the typical algorithms of each type of method.