Palette based colour transfer using image segmentation

While the ability to see colours is innate, their perception is a guided process. Varying across languages and cultures, linguistic vocabulary is one such influence that guides our perception. In colour specifically, the names of colours in a language act as verbal cues that help us memorize and later recognize colours of the spectrum. This makes colour a cultural and local phenomenon. This paper presents a novel method to perform colour transfer, given an image and an elaborate colour palette as inputs. Image segmentation is employed, followed by a region based colour transfer to ensure accurate local colour mapping. The method is used to bias the given image with colours that are perceptually closest to it in a reference language. This is done in an attempt to make the colour transfer reflect the changes in colour vocabularies of people from different languages and cultures.