TY - JOUR
AU - Ziwei Zhang
AU - Zhenzhen Wan
AU - Haotian Zhu
AB - As an important component of human civilization, intangible cultural heritage (ICH) faces many challenges in its protection and inheritance. The purpose of this article is to propose an intelligent algorithm based on a generative adversarial network (GAN), which can be used to extract core visual elements from ICH images efficiently and apply it to modern graphic design. In this study, a GAN model combining U-Net structure and the CycleGAN module was constructed, and the precise extraction of ICH visual elements under complex background was realized by the antagonistic learning mechanism of generator and discriminator. Experiments showed that the algorithm is superior to the traditional methods in terms of intersection over union, peak signal-to-noise ratio, and structural similarity index, reaching 85.6\%, 31.5 decibels, and 0.91, respectively, which significantly improved the extraction accuracy. The user feedback survey showed that more than 85\% of designers thought that the extracted visual elements had high cultural identity and aesthetics.
DO - 10.4018/IJITSA.384915
M1 - 1
N2 - As an important component of human civilization, intangible cultural heritage (ICH) faces many challenges in its protection and inheritance. The purpose of this article is to propose an intelligent algorithm based on a generative adversarial network (GAN), which can be used to extract core visual elements from ICH images efficiently and apply it to modern graphic design. In this study, a GAN model combining U-Net structure and the CycleGAN module was constructed, and the precise extraction of ICH visual elements under complex background was realized by the antagonistic learning mechanism of generator and discriminator. Experiments showed that the algorithm is superior to the traditional methods in terms of intersection over union, peak signal-to-noise ratio, and structural similarity index, reaching 85.6\%, 31.5 decibels, and 0.91, respectively, which significantly improved the extraction accuracy. The user feedback survey showed that more than 85\% of designers thought that the extracted visual elements had high cultural identity and aesthetics.
TI - Intelligent Extraction of Intangible Cultural Heritage Visual Elements and Application in Graphic Design Based on Generative Adversarial Network
VL - 18
SN - 1935-570X
ER -