The Role of Artificial Intelligence Algorithms in Personalized Design of Visual Communication

Based on the problems of insufficient design innovation and low design efficiency in personalized visual communication design, this study uses the style transfer algorithm to apply a specific artistic style to the design and automatically generates visual elements with personalized characteristics. At the same time, Generative Adversarial Networks (GAN) are used to create new patterns and color combinations and enhance the diversity of design, while ChatGPT is used to parse user needs and transform text descriptions into design concepts. In the process of implementing these methods, this paper collects and analyzes user data to build user portraits to understand their preferences and needs. This paper also uses a style transfer algorithm to combine user-favorite elements with design templates to generate preliminary design sketches. Then GAN intervenes to generate new design elements and uses these elements to enrich and refine the design content. During the design process, the artificial intelligence model is continuously optimized based on user feedback to ensure that the final output design not only meets the user's personalized needs but is also highly innovative and attractive. Ultimately, these individually adjusted designs are output in an appropriate format for application in actual products or services. Based on comparative experiments, this paper found that personalized design is superior to traditional design methods in multiple key indicators: the average innovation score of personalized design is as high as 9.06, higher than the 6.89 of traditional design; the time required for personalized design is mainly concentrated between 20 and 30 minutes, while the time required for traditional design ranges from 30 to nearly 50 minutes and fluctuates greatly; the adoption rate of personalized design reaches a minimum of 90.2% and a maximum of 99.9%, while the adoption rate of traditional design is lower, with a minimum of only 75.2% and a maximum of only 88.7%. Through comparative analysis, this paper confirms the advantages of personalized design over traditional design in multiple dimensions, and demonstrates the application potential and value of artificial intelligence technology in the field of design.