| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide an image processor and method that can generate a three-dimensional model from a live video image of an event, such as a soccer match or the like, can change a view of the three-dimensional model, and can provide a view from a view point at which no camera is present in a real world. <P>SOLUTION: A three-dimensional representation of a scene including a plurality of objects disposed on a plane is generated. The three-dimensional representation is generated, by extracting image features of each of the objects from a view of the scene generated from one or more video images, comparing the image features with sample image features of assumed objects, and identifying the objects. Further, the three-dimensional representation can be generated, by generating object path data that indicate a position of the objects on the plane in the video images with respect to time, calculating a projection matrix for projecting the position of the objects from the plane in the video images into a three-dimensional model, and projecting the position of the objects on the basis of the object path data into the plane of the three-dimensional model of the scene using the projection matrix and a predetermined estimated value of the height of each of the objects. <P>COPYRIGHT: (C)2009,JPO&INPIT |
