Research And Implementation Af Intelligent Generation Method Of Ceramic Decorative Patterns

Jingdezhen is a city famous for its ceramics,and the divine fire of ancient kilns is still bright for thousands of years.The shape and decorative patterns in ceramics are one of the main research contents of the researchers.However,the decorative patterns are timeconsuming and labor-intensive in designing the patterns,and the traditional decorative design style is single and the concept is outdated,which cannot meet the modern aesthetic requirements.Aiming at the limited versatility of the generation method of ceramic decorative patterns,this paper proposes a method to generate ceramic decorative patterns using intelligent algorithms.This paper first introduces the background and significance of the subject research,summarizes the research status and development trends at home and abroad,and expounds the related research methods and implementation principles of image style transfer.The local migration algorithm of the fused ceramic decorative pattern is discussed.Finally,the method of intelligent generation of ceramic decorative pattern is discussed by synthesizing a variety of network structures.The traditional image style transfer algorithm ignores the edge distribution of the image,which makes the contour of the generated image blurry and takes a long time in the iterative process of style transfer.Therefore,this paper proposes a style transfer algorithm for ceramic decorative patterns based on the ESPCN model.The algorithm uses the Laplace operator to sharpen the image to highlight the edge distribution,uses downsampling to generate low-resolution images to reduce the iteration time of image style transfer,and then uses ESPCN super-resolution reconstruction to convert lowresolution images into high-resolution images image.The experimental results show that the generated image has a clear edge distribution,shortened iteration time,and improved image quality and clarity through multiple image evaluation indicators.In addition,in view of the limitation of image style transfer local area,this paper proposes a local transfer model of ceramic decorative patterns based on Poisson fusion.The model mainly transfers content images and style images locally,and uses techniques such as semantic segmentation and Poisson fusion to complete the transfer task.Experiments show that the model can effectively perform local migration,and the generated decorative images can be well combined with ceramics,showing the perfect combination of art and technology.Finally,in view of the large fluctuation phenomenon of DCGAN in the training of loss function during the training process,it is improved,and the four network structures of GANs are intelligently generated for ceramic decorative patterns.The intelligent generation method of ceramic decorative patterns is mainly to use different models to intelligently generate decorative patterns,thereby reducing costs,saving time,and completing the generation of decorative patterns more intelligently.This method can be applied to many smart scenes,so that in the production process of ceramics,the content of decorative patterns can be more personalized,intelligently designed and displayed,and the decorative patterns required in the ceramic type can be efficiently generated.The intelligent design of ceramic products provides technical support and method guidance.