Cartooning using reference images

Abstract

This paper proposes the method that makes an input target image into exaggerated cartoon-like images by using reference images. To deform a target image, we extract feature points from a target image and define the feature point model on reference images. And then, we apply feature based warping method to this deformation. For our result be felt more cartoonish, we additionally apply the luminance quantization method and the edge enhancement method to the deformed target image. At this time, we control intensities of the target image deformation, the luminance quantization and the edge enhancement for the capability that is able to create various results. © 2010 IEEE.