Automatic generation of traditional patterns and aesthetic quality evaluation technology

As an significant element for the continuation and development of modern design, conventional Chinese decorative patterns satisfy people's emotional dependence on nationality tradition. At present, traditional are designed mainly by professional designers. Due to high entire design process labor, efficiency is untoward improve rapidly meet explosive growth business needs. However, a large amount material data Internet artificial intelligence technology provides opportunity work out this bottleneck. This paper studies implementation automatic generation pattern image layout based confrontation network aesthetic evaluation score quality generated image. And scoring results fed back pictures so that structure can adapt disparate scene requirements. The show SSIM value structural similarity obtained automatically model in larger closer 1 than other methods, indicating has highest quality. Furthermore, research assistance intelligent solve bottleneck design. Finally, future digital approach prospected.