Novel approach to neural network based caricature generation

Authors:

Abstract
A caricature is defined as a funny drawing of someone that makes some of his/her distinct features appear exaggerated or more amusing. However the caricatures of the same person created by different artists can be very different, since the artists drawing styles play an important role (Hughes, 1999). Therefore learning the drawing style of an artist provides the key to the computer based automatic generation of professional caricature. Unfortunately, no caricature generation system in the past has attempted to address this issue with the aid of artificial intelligence technologies. In this paper, the authors propose an example-based caricature generation system with experimental results and detailed analysis to prove that neural networks can be used for capturing the drawing style of an artist. This work is the first system to use neural networks in generating caricature.