Example-based makeup transfer

As one of the significant applications of image transformation, digital makeup combines efficiency and entertainment, providing inspiration for the development of computer vision in the fields of fashion, film, and games. However, the traditional makeup transformation method is time-consuming and requires a pair of “before” and “after” makeup images as examples. Also, how to separate the information from the illumination and color in the images is another issue. Therefore, in this paper, we propose an example-based makeup transfer algorithm for image composition and re-composition by using facial landmark, face warping in adjusted MLS, fast guided filter, and alpha-blending methods. Experiments demonstrate that our method has higher speed in image decomposition. More importantly, we find fitting parameters that have a better impact on color transfer in image composition.