| 摘要 |
<p>A method for describing the contents of dynamic image data efficiently with a small amount of data, including a mechanism in which (1) the trace along which each object moves with time is expressed by using a reference plane for expressing the information on the position of the object, (2) a description unit based on the type of the motion of the object is provided by using the change of the shape of the object, (3) the motion of the object is represented as a motion section, and (4) the definition of the object depending on the video contents, the definitions of the action types, the definitions of the scene interpretation depending on the interaction of the objects are read and interpreted.</p> |
