On a university campus there are a variety of cultural resources. However, the lack of services to display such items often causes them to go unnoticed. Moreover, in Japan visual education has not yet been put into practice. To overcome this problem we have developed a website for the purpose of leading users to cultural resources through visuals. Using pictures taken by site visitors, our system guides them to resources located throughout the campus. The pictures are used as queries for a Content-Based Image Retrieval system (CBIR) from which other images similar in shape and colour are returned as results. This paper argues that one way to develop visual literacy is to recognize the relation between physical objects and landscape. To better understand this relationship, this paper examines how we look at objects with reference to computer science. To that end, extraction methods from image retrieval systems are considered. The process of our system reduces information from images, initially removing from them features such as depth, surface orientation, etc., or by replacing colour ranges. Users compare images acquired through these methods with real physical objects and recognize their diversity. Because each image branches out into other similar images, the diversity will lead to a tree-structured image world. This paper concludes that an open-end image retrieval system would provide an interactive platform for the expansion of visual literacy.