基于交互式遗传算法的参考图像辅助配色设计

Reference Image Aided Color Matching Design Based on Interactive Genetic Algorithm

目的降低配色过程中意象表达的不稳定性,在配色方案中,辅助设计师将隐性的设计理念表达出来。方法设计师使用心仪图像作为配色意象的参考,并采用交互式遗传算法开发辅助设计师的配色技术,将提取的各种色彩组合映射至配色对象,供设计师评价与选择。在CorelDraw平台上搭建原型系统,设计人机交互界面。以敦煌壁画为配色意象来源,进行图像配色的实例分析。结果实现了方案的自动批量生成及其概率选择、交叉重组、变异、色区调整等系列遗传操作,得到了与图像意象融合的最优配色方案。结论在色彩提取值框架内,设计师通过交互式遗传算法辅助软件,从全局视角观察配色方案,在人机互动作用下,逐步优化方案,以便在短时间内搜索到符合设计师配色思维的方案。

The work aims to reduce the instability of image expression during color matching and assist designers in expressing the hidden design concept in the color scheme. With the designer’s favorite image as a reference for the color matching image, the interactive genetic algorithm was used to develop the technology to assist the designer. Various extracted color combinations were mapped to color matching objects for designer evaluation and selection. The prototype system was built on the CorelDraw platform and the human-computer interaction interface was designed. The Dunhuang murals were taken as color matching image source to carry out the case analysis on the color matching of images. The automatic mass production of the scheme and its genetic operations such as probability selection, cross recombination, mutation, and color zone adjustment were realized to obtain the optimal color scheme embodying the imagery of reference image. Within the framework of the color extraction value, the designer can observe the color scheme from a global perspective through interactive genetic algorithm assistant software, and gradually optimize the scheme under the interaction of human and computer, so as to find the scheme in line with the color matching thought of the designer in a short time.