Enhancing the impact of color through artificial intelligence for visual narration

Youngsue Han(Author, Part-time lecturer, Sogang University), Hyunjung Roh(First co-author, Teacher, Chungang Primary School), Jaehoon Kim(Second co-author, Graduate student, Sogang University), Suhyoun Hwang(Third co-author, Graduate student, Sogang University) 

In the seventeenth century, Isaac Newton claimed that white light is heterogeneous and consists of multiple component colors. Newton also stated that color is a mechanical and measurable attribute. A century later, Johann Goethe conducted a series of experiments for analyzing the nature of color. In contrast to Newton’s theory, Goethe postulated that the color theory should address the subjective and psychological impacts of color on human beings. Goethe’s principle has significantly influenced the contemporary philosophical and artistic approaches to the color theory. In this study, based on Goethe’s color theory, the authors developed separate color categories for the film “La La Land” and incorporated machine learning algorithms. To achieve the said objective, structural analysis was conducted by employing the “five code network or topos” described in Roland Barthes’s book titled S/Z that was published in the year 1970 . The Barthesian code network was thereby applied to create input training data for machine learning algorithms. The proposed study thus comprises foundational and novel research that can aid in developing an artificial intelligence framework for classifying the psychological responses of viewers to colors.

Key words：Artificial Intelligence, Psychology of Color, Goethe, Barthesian Code Network, Classification, Machine Learning