Digital image colorization based on probabilistic distance transformation

Colorization is a term used to describe a computerized process for adding color to black and white pictures, movies or TV programs. This process replaces a scalar value that represents pixelspsila intensity with a vector in a three dimensional color space composed of hue, saturation and luminance, and since this mapping between intensity and color has no inherently rdquocorrectrdquo solution, human interaction and/or external information usually plays a large role. In this paper we present a novel colorization method that takes advantage of the modified morphological distance transform to automatically propagate the color scribbled by a user on the gray scale image. The introduced modification of the distance transform is based on the Gibbs distribution, which governs the behavior of a virtual particle performing a random walk on the image domain. Such a modification allows for the application of the distance transform to the gray scale images and yields high quality colorization results.