Intelligent Design System For Personalized Ceramic Products

With the rapid development of artificial intelligence as well as industrial technology,modern manufacturing has also emerged.The traditional ceramic industry is lagging behind.With the emergence of 3D printing technology,it is now possible to achieve intelligent ceramic production,but there is less research on the intelligent design of personalized ceramic products.Due to the wide variety of ceramics and their complicated names,and the incomplete understanding of ceramic shapes and ceramic manufacturing technology by non-professionals,there are problems of long and difficult design cycles.To address these problems,this project researches and develops an intelligent design system for personalized ceramic products that meets multimodal needs,enabling users’ personalized needs to be realized through artificial intelligence technology,and injecting new vitality into the development of the ceramic industry.In this thesis,based on the research of theoretical knowledge of natural language,computer vision and other technologies,firstly,a multimodal demand acquisition method is proposed for multiple personalized demands of users.The method mainly involves: 1)text-based and speech-based personalized demand acquisition through Text Rank-based keyword extraction technology;2)accurate classification of ceramic images through an improved Alex Net-based ceramic image classification method,the classification result is the most direct expression of the user’s demand expressed through the image,and the classification result can be used to know what kind of porcelain the user needs,so that the The classification result can be used to know what kind of porcelain the user needs as the user’s personalized demand;3)through the ceramic video detection method based on YOLOv7,the video frames containing ceramic images in the user’s uploaded video are identified,so as to obtain the user’s demand expressed through the video.Then,multiple demand acquisition methods are combined to achieve the acquisition of multimodal personalized design demands.Based on this,a ceramic 3D model retrieval based on multimodal demand acquisition is proposed,which efficiently achieves 3D ceramic model retrieval by performing similarity calculation between the acquired multimodal demands and model annotation.Finally,the 3D model modeling of 54 types of ceramics was carried out using 3Dmax,and a personalized ceramic product intelligent design system was developed based on Vue and Three.js and python.Through this system,users can freely select ceramic models,decals and base models to achieve intelligent combinations,and can realize the intelligent design of personalized ceramic products through multimodal requirements such as voice,image and video.The personalized ceramic product intelligent design system studied and imple mented in this paper is an application of artificial intelligence technology to cera mic product design.The system is based on multimodal demand acquisition for r etrieval of 3D ceramic models,and on the basis of the retrieved ceramic models,the intelligent design of ceramic products is realized by automatic combination te chnology.The system can improve the efficiency of ceramic product design and h as positive significance to promote the development of ceramic industry.