| 摘要 |
<P>PROBLEM TO BE SOLVED: To quickly detect a face image portion even when a face moves. <P>SOLUTION: An estimated position of the face image portion Fn in a subject image in the n-th frame is predicted based on a position of the face image portion F(n-2) detected from the subject image in the (n-2)-th frame and a position of the face image portion F(n-1) detected from the subject image in the (n-1)-th frame. A target area An is set to detect the face image portion in the subject image in the n-th frame, based on the predicted position. The face image portion is detected in the subject images in the n-pieces of frame, within the set target area An. The face image portion is relatively quickly and accurately detected because the face image portion is detected within the set area set by predicting the motion of the face image portion. <P>COPYRIGHT: (C)2008,JPO&INPIT |
