The dot diffusion method for digital halftoning has the advantage of parallelism unlike the error diffusion method. The method was improved by optimization of the so-called class matrix so that the resulting halftones are comparable to the error diffused halftones. First, 8 × 8 class matrices were used for dot diffusion method. However, there is a problem with this size of class matrix: enhancement of images is necessary before halftoning. It was found later that if the size of the class matrix was increased to 16 × 16, then there is no need for the enhancement step. In this article, we will review the dot diffusion. This will be followed by a discussion on special cases of dot diffusion. Then, we will show how the optimization is done to get the class matrices.