Visitor-centered artwork network analysis using object detection

Abstract

The recent museum exhibition is turning into a way of satisfying visitors' demand. In order to provide the visitor-centric exhibit, the artwork needs to be analyzed based on visitor's behavior, not only by the museum professional's point of view. This study aims to analyze the relation between artworks concerning network analysis based on the visitor's behavior using object detection techniques. The cameras installed in a museum recorded visitors, and the object detector and the content-based image retrieval technique tracked visitors from videos. The time spent on the artwork was measured, and the data was converted into a bipartite graph. The relationship between artworks was analyzed with the visitor-centered artwork network. Based on the visitor's behavior, the significant artworks were identified, and the artwork network was compared to the arrangement of a museum. The tendency of edges in artwork network is also examined with visitors' preference for artworks. The method of this study makes it possible to collect quantitative data, and the result could be used as a basis and reference for analyzing the artwork in a visitor-centered approach.