Feasibility Study of Non-Photorealistic Computer Automatic Stylization for Chinese Cartoon Painting (Early Manhua Style)

Our research project aims to participate in a long-term collaboration between computer scientists and artists to develop a non-photorealistic graphics tool capable of simulating Chinese artistic styles. At the initial feasibility study stage, we chose a subset of the work of Feng Zikai as a case study. This paper presents the preliminary results of our prototype system, which attempts to automatically stylize an input portrait image, as well as video, into Chinese cartoon painting style.Feng Zikai (1898-1975) was an outstanding modern Chinese cartoonist as well as an accomplished essayist, calligrapher, scholar, musician, translator, and educator. He was chiefly known for developing a unique style of cartoon painting (manhua, in Chinese) incorporating a combination of traditional Chinese fine arts features with references to Western painting and drawing styles.This paper discusses two phases: the first involves a systemic study of the sources, methods, and techniques of Feng Zikai’s cartoon paintings; and the second is to build a computer program capable of simulating Feng’s styles.Aside from the potential technical benefits of such study, a comprehensive understanding of this Chinese cartoon painting style may also inspire a discussion of the relationship between tradition and novelty in the world of art.