This paper proposes a new method which simplifies shapes of objects in the real world by giving them their meanings so that a real-world-scene database can be realized. To construct a database for shapes and movements of objects in a world scene, it is necessary to efficiently represent the shape and meaning of each object by using a small amount of data, and to establish a query process relating to them. In the proposed method, each scene is represented by an approximate polygon model (called shape of the individual model) which is based on the shape and meaning of each object. This database is constructed by mapping shapes of individual models and their three-dimensional movements (extracted from a scene) to a virtual CG space. The shapes of individual models have actually been made, and their accuracy, and realization of the query process based on the space operator have been evaluated. © 2000 Scripta Technica, Syst Comp Jpn, 31(12): 6272, 2000

This paper proposes a new method which simplifies shapes of objects in the real world by giving them their meanings so that a real-world-scene database can be realized. To construct a database for shapes and movements of objects in a world scene, it is necessary to efficiently represent the shape and meaning of each object by using a small amount of data, and to establish a query process relating to them. In the proposed method, each scene is represented by an approximate polygon model (called shape of the individual model) which is based on the shape and meaning of each object. This database is constructed by mapping shapes of individual models and their three-dimensional movements (extracted from a scene) to a virtual CG space. The shapes of individual models have actually been made, and their accuracy, and realization of the query process based on the space operator have been evaluated. © 2000 Scripta Technica, Syst Comp Jpn, 31(12): 6272, 2000