Techasarntikul, NattaonRatsamee, PhotcharaOrlosky, JasonMashita, TomohiroUranishi, YukiKiyokawa, KiyoshiTakemura, HaruoKakehi, Yasuaki and Hiyama, Atsushi2019-09-112019-09-112019978-3-03868-083-31727-530Xhttps://doi.org/10.2312/egve.20191276https://diglib.eg.org:443/handle/10.2312/egve20191276Augmented Reality (AR) has become a popular technology in museums, and many venues now provide AR applications inside gallery spaces. To improve museum tour experiences, we have developed an embodied agent AR guide system that aims to explain multi-section detailed information hidden in the painting. In this paper, we investigate the effect of different types of guiding interfaces that use this type of embodied agent when explaining large scale artwork. Our interfaces include two types of guiding positions: inside and outside the artwork area, and two types of agent movements: teleporting and flying. To test these interfaces, we conducted a within-subjects experiment to test Inside-Teleport, Inside-Flying, Outside-Teleport, and Outside- Flying with 28 participants. Results indicated that although the Inside-Flying interface often obstructed the painting, most of the participants preferred this type since it was perceived as natural and helped users find corresponding art details more easily.Humancentered computingHCI design and evaluation methodsComputing methodologiesMixed / augmented reality10.2312/egve.2019127625-32