Anisotropic Lp Poisson Disk Sampling for NPR Image with Adaptively Shaped Pieces

Abstract

A non-photorealistic rendering technique is presented for generating images such as stippling images and paper mosaic images with various shapes of paper pieces. Paper pieces are spatially arranged by using an anisotropic L_p poisson disk sampling. The shape of paper pieces is adaptively varied by changing the value of p. We demonstrate with experiments that edges and details in an input image are preserved by the pieces according to the anisotropy of their shape.