The Application of Semantic Network Theory to Visual Art

Complexity science uses complex networks to algorithmically examine many-body systems containing varying types of interactions. Though commonly used in STEM applications, this form of modeling is effective when used to analyze the arts and humanities. This project applies complex network theory to visual arts in pursuit of a computational system that can determine the artistic movement any arbitrary 2D piece originates from. The system will interrogate the visual characteristics of pieces from the Renaissance period of the 1400s through to the Surrealist period of the 1950s. Following a semantic network blueprint, the system takes an input image, reduces that image into a collection of super pixels, builds a network, and performs network measurements. These measurements are recorded, and after all the images have been analyzed, the different artistic movements are distinguished from each other through a statistical analysis. By quantitatively distinguishing between art periods, the system will draw qualitative conclusions on the evolution of art, such as the motivation in stylistic change.