Digital Modeling of Intangible Cultural Heritage Visualization Based on Convolutional Neural Networks

Abstract

This article introduces an innovative digital modelling method tailored for ICH using the functionality of Convolutional Neural Networks (CNN). In order to further improve the performance of CNN, we seamlessly integrated a genetic algorithm (GA) into the system, thus establishing an immersive interactive visual learning environment. This integration utilizes GA's proficiency in conducting comprehensive global searches and its ability to optimize through genetic operations such as crossover and mutation iteratively. The parameters of CNN are fine-tuned based on the fitness value, which represents the most accurate solution with the smallest deviation. This study has achieved remarkable results. Compared with traditional modelling techniques, CAD technology has shown significant advantages in modelling accuracy. In terms of visual teaching, CAD technology enables students to intuitively understand the structure and characteristics of intangible cultural heritage by creating realistic 3D models. This intuitive teaching method not only stimulates students' interest in learning but also helps them gain higher cultural and material value. This method helps to promote a deeper appreciation and understanding of our rich cultural heritage among the younger generation.