A deep learning-based method for evaluating the quality of corporate brand packaging visual design

Abstract

Abstract With the development of deep learning technology, the quality evaluation of enterprise brand packaging visual design becomes more critical. The study first established a brand packaging design product color imagery dataset through color emotionalization, and used systematic clustering technology for imagery selection and evaluation. Subsequently, the brand packaging visual design was optimized based on user demand, combining perceptual engineering and user demand mapping model. Many samples were evaluated by GoogLeNet model, and the data were processed by K-mean clustering and semantic difference method. The results show that the proposed method can effectively distinguish the perceptual imagery of different brand packaging designs, such as traditional, modern, simple, and complex. Specifically, more than 90% of the samples in the experiment achieve high consistency in perceptual imagery evaluation. In addition, the study analyzed the classification effect and quality evaluation of corporate brand packaging visual design, proving the validity and reliability of the method. This study provides a new quality evaluation method for corporate brand packaging graphic design, which helps to improve design efficiency and quality.