Research on AIGC-driven Innovation Path and Smart Creative Workflow Optimization in Design Field

AIGC has attracted much attention in the application of art and design professional fields and has gradually become an important auxiliary tool for art and design professionals creation. This paper utilizes AIGC as a starting point for design creativity and develops a text-guided image generation model based on AIGC, which employs a semantically consistent generative adversarial network. We compare different model indexes across different datasets to evaluate the image generation effect of the constructed model. Subsequently, we summarize and optimize the intelligent, creative workflow. Based on this workflow and the image generation model, we generate a series of images as an example of a cosmetic package. We then analyze the audience’s acceptance of these intelligently generated images through a questionnaire survey, thereby exploring the impact of AIGC-driven design and innovation. The GAN-SC model outperforms the comparison algorithms in various datasets, achieving an IS value higher than AttnGAN and DFGAN by 17.7% and 2.7%, respectively, and reducing the FID metrics by 29.43% and 19.6%, respectively. The acceptance rate of the AIGC-based design solution surpasses that of the traditional solution by 4.9%, and it excels across all design dimensions, showcasing the effectiveness of AIGC-driven design innovation.