(A) scene graph-based facing-formation seat matching method for generating a collaboration space between heterogeneous spaces

Abstract

This study proposes a scene graph-based facing-formation seat matching method for maximizing the number of commonly available seats(matching seats) when generating collaboration space between heterogeneous spaces. collaboration space is necessary for participants to meet and interact with each other. However, dissimilarity between spaces was a challenge for the researchers when generating the collaboration space. Most of the previous studies have studied solutions using empty areas in the physical space, and also aimed to maximize the size of the space. This study aims to maximize the matching seats by focusing on the desk-chair scene consisting of a seat and a work space in the space for the scenario of sitting and collaborating. Our method finds seats belonging to the same face-to-face seating formation based on the positional relationship between two affordance areas, the area where users can sit and the area where they can work. Through the experiment, the ratio of matching seats among seats in the space (seat matching rate) was compared with the existing method. As a result of experiments with the dissimilarity between spaces defined by the Manhattan distance and the generated test dataset, the seat matching rate decreases as the dissimilarity increases, but the proposed method shows a higher seat matching rate than the existing methods. This study shows the possibility of conducting a collaboration scenario using chairs in physical space between heterogeneous spaces with high dissimilarity.