Pixelated Image Abstractoin

What makes pixel art both compelling and difficult is the limitations imposed on the medium. With a significantly limited palette and resolution to work with, the task of creating pixel art becomes one of carefully choosing the set of colors and placing each pixel such that the final image best depicts the original subject. This task is particularly difficult as pixel art is typically viewed at a distance where the pixel grid is clearly visible, which has been shown to contribute to the perception of the image.

We present an automatic method that can be used to abstract high resolution images into very low resolution outputs with reduced color palettes in the style of pixel art. Our method simultaneously solves for a mapping of features and a reduced palette needed to construct the output image. The results are an approximation to the results generated by pixel artists. We compare our method against the results of a naive process common to image manipulation programs.

In the first part of each iteration we use a modified version of an image segmentation proposed by Achanta et al. [2010] to map regions of the input image to output pixels. In the second step, we utilize an adaptation of mass-constrained deterministic annealing [Rose1998] to find an optimal palette and its association to output pixels. These steps are interdependent, and the final solution is an optimization of both the physical and palette sizes specified by the user.