Graphics, Imaging, and Games Lab

In this paper, we present an improvement to the minimum error boundary cut, a method of shaping texture patches for nonparametric texture synthesis from example algorithms such as Efros and Freeman’s Image Quilting. Our method uses an alternate distance metric for Dijkstra’s algorithm, and as a result we are able to prevent the path from taking shortcuts through high cost areas, as can sometimes be seen in traditional image quilting. The resulting artifacts tend to be less pronounced than those created using the traditional cumulative distance metric. Post-process methods such as pixel re-synthesis can easily be modified and applied to our minimum error boundary cut to increase the quality of the results. This technique is not only useful for texture synthesis, but also seems promising for applications such as Wang Tiles or other contexts in which the minimum error boundary cut is applied.