Abstract: Colorization of grayscale image is a challenging research field in image processing, and has wide application prospect. After analyzing current colorization algorithms, the paper proposes a new colorization algorithm based on p-Laplace equation. A user needs to provide the grayscale image with a few color scribbles manually. Then, the grayscale image is colorized by solving the p-Laplace equation. Since the p-Laplace operator is a nonlinear anisotropic diffusion operator, the new method can better maintain the color’s effect of the edge, at the same time achieving the color expansion than Saprio’s poisson colorization algorithm and Teng sheng-hua’s Laplace colorization algorithm, which are also partial differential equation based methods.

Key words: image colorization, p-Laplace equation, partial differential method, image processing