It is intended to solve the problem that, when an operator performs an operation using two-dimensional input means while seeing a picture frame seen from a viewpoint of the operator, an input operation is difficult to perform when the direction of the line of sight of the operator is unstable, since a designated direction greatly deviates in a world coordinate system. For that purpose, viewpoint information is detected, and an instruction of the operator for operating a position of a pointer image is input. A designated direction in a pointer coordinate system is obtained in accordance with the operator's instruction, and the pointer image is generated based on the designated direction. The pointer coordinate system is changed from the detected viewpoint information in accordance with a specific instruction of the operator.