Color Transfer between Images with IEC

Qiang Li¹, Tsuyoshi Nakamura¹*, Yuyan Chao^2,4, Lifeng He^3,4 and Hidenori Itoh¹

¹Department of Computer Science and Engineering, Nagoya Institute of Technology, Gokiso-cho, Showa-ku, Nagoya 466-8555, Japan
²Faculty of Environment, Information and Business, Nagoya Sangyo University, 3255-5 Arai-cho, Owariasahi, Aichi 488-8711, Japan
³Faculty of Information Science and Technology, Aichi Prefectural University, 1522-3 Ibaragabasama, Kumabari, Nagakute-cho, Aichi-gun, Aichi 480-1198, Japan
⁴Guest Professor of Shanxi University of Science and Technology, China
*E-mail address: tnaka@ics.nitech.ac.jp

(Received December 1, 2004; Accepted January 31, 2005)

Keywords: Non-photorealistic Rendering, Color Transfer, Interactive Evolutionary Computation

Abstract. Some research has been conducted on applying coloration of painterly art to a photographic image. Most of the proposed methods use some absolute criteria, e.g., the nearest distance in CIE L*a*b* color space, converting a color to another uniquely. However, we think that basic color transfer depends on the subjective feelings and sensitivities of users themselves. We describe an algorithm where every user can produce an original color-transformed image resulting from two inputs, which are a target photograph and a reference painting. Our algorithm is based on interactive evolutionary computation (IEC) and can produce a variety of color-transformed images semi-automatically. Some color-transformed examples similar to reference paintings are illustrated, and the searching efficiency of our method is discussed. Figures 1, 2, 3, 4, 13, 14 and 15 are originally made of color images. They can be seen in WEB version of FORMA. URL is "http://www.scipress.org/jounals/forma/index.html."