| 摘要 |
PROBLEM TO BE SOLVED: To solve problems such that there is a large error when a speaker is detected only by his voice, and there is a large amount of computation because of pursuing the contour in the image by the technique suggested as a method for correcting with image after having been detected by the voice roughly in a video conference terminal. SOLUTION: The system comprises a detector part 3 for detecting the direction of a speaker using phase difference in voice signal supplying to multi- microphones 1 and a control part 7 by which the camera 2 is pointed in the detected direction. A moving part is detected by using frame difference at motion detecting part 4 from video images picked up with the camera pointing in the direction of speaker's voice and the distribution is measured at the motion distribution measuring part 5 and then, the personal position is detected by the motion distribution in the personal position detecting part 6. Even if the speaker is not in the center position of taken image, an appropriate speaker's image enables to take by moving the camera with the camera control part, such as a detected person is in the center of the camera. |
