Interactive Design of Dynamic Visual Communication Driven by Artificial Intelligence Algorithms

Dynamic visual communication design usually needs to process multimodal information. How to effectively integrate these multi-dimensional features and establish stable associations between different modalities remains a technical challenge. To this end, this article explores the application of artificial intelligence algorithms, especially spatial attention and channel attention mechanisms, in dynamic image and video generation to achieve more accurate semantic consistency between text and images. This article introduces a spatial attention mechanism, which dynamically focuses on the spatial area related to the text description by calculating the similarity between text and image features; at the same time, the channel attention mechanism further optimizes the match with the text description by adjusting the channel weights in the image feature map. In addition, the study introduces user interaction design, and users can influence the generation target of the generator through real-time feedback and adjustment, making it more personalized and in line with needs. The optimization process of the generator and the discriminator further improves the realism and quality of dynamic image and video generation through GAN (Generative Adversarial Network). Experiments show that the proposed method can effectively enhance the semantic consistency in the image generation process and provide new technical support for dynamic visual communication design.